3 * See https://wiki.hpdd.intel.com/display/PUB/Lustre+Support+Matrix
4 for currently supported client and server kernel versions.
5 * Server known to build on patched kernels:
6 2.6.32-431.29.2.el6 (RHEL6.5)
7 2.6.32-504.16.2.el6 (RHEL6.6)
8 3.10.0-229.1.2.el7 (RHEL7.1)
9 3.0.101-0.47.52 (SLES11 SP3)
10 * Client known to build on unpatched kernels:
11 2.6.32-431.29.2.el6 (RHEL6.5)
12 2.6.32-504.16.2.el6 (RHEL6.6)
13 3.10.0-229.1.2.el7 (RHEL7.1)
14 3.0.101-0.47.52 (SLES11 SP3)
15 * Recommended e2fsprogs version: 1.42.9.wc1 or newer
16 * NFS export disabled when stack size < 8192 (32-bit Lustre clients),
17 since the NFSv4 export of Lustre filesystem with 4K stack may cause a
18 stack overflow. For more information, please refer to bugzilla 17630.
19 * NFSv4 reexport to 32-bit NFS client nodes requires Lustre client on
20 the re-exporting nodes to be mounted with "32bitapi" mount option
22 --------------------------------------------------------------------------------
24 03-10-2015 Intel Corporation
26 * See https://wiki.hpdd.intel.com/display/PUB/Lustre+Support+Matrix
27 for currently supported client and server kernel versions.
28 * Server known to build on patched kernels:
29 2.6.32-431.29.2.el6 (RHEL6.5)
30 2.6.32-504.8.1.el6 (RHEL6.6)
31 3.0.101-0.46 (SLES11 SP3)
32 * Client known to build on unpatched kernels:
33 2.6.32-431.29.2.el6 (RHEL6.5)
34 2.6.32-504.8.1.el6 (RHEL6.6)
35 3.10.0-123.20.1.el7 (RHEL7)
36 3.0.101-0.46 (SLES11 SP3)
37 * Recommended e2fsprogs version: 1.42.9.wc1 or newer
38 * NFS export disabled when stack size < 8192 (32-bit Lustre clients),
39 since the NFSv4 export of Lustre filesystem with 4K stack may cause a
40 stack overflow. For more information, please refer to bugzilla 17630.
41 * NFSv4 reexport to 32-bit NFS client nodes requires Lustre client on
42 the re-exporting nodes to be mounted with "32bitapi" mount option
44 Severity : enhancement
46 Description: control OST-index in IDIF via ROCOMPAT flag.
47 Details : Introduce new flag OBD_ROCOMPAT_IDX_IN_IDIF that is stored in the
48 last_rcvd file. For new formatted OST device, it will be auto set;
49 for the case of upgrading from old OST device, you can enable it
50 via the lproc interface osd-ldiskfs.index_in_idif. With such flag
51 enabled, for new created OST-object, its IDIF-in-LMA will contain
52 the OST-index; for the existing OST-object, the OSD will convert
53 old format IDIF as new format IDIF with OST-index stored in the
54 LMA EA when accessing such OST-object or via OI scrub. Once such
55 flag is enabled, it cannot be reverted back, so the system cannot
56 be downgraded to the orignal incompatible version.
58 --------------------------------------------------------------------------------
60 07-30-2014 Intel Corporation
62 * See https://wiki.hpdd.intel.com/display/PUB/Lustre+Support+Matrix
63 for currently supported client and server kernel versions.
64 * Server known to build on patched kernels:
65 2.6.32-431.20.3.el6 (RHEL6.5)
66 3.0.101-0.31 (SLES11 SP3)
67 * Client known to build on unpatched kernels:
68 2.6.32-431.20.3.el6 (RHEL6.5)
69 3.0.101-0.31 (SLES11 SP3)
72 * Recommended e2fsprogs version: 1.42.9.wc1 or newer
73 * NFS export disabled when stack size < 8192 (32-bit Lustre clients),
74 since the NFSv4 export of Lustre filesystem with 4K stack may cause a
75 stack overflow. For more information, please refer to bugzilla 17630.
76 * NFSv4 reexport to 32-bit NFS client nodes requires Lustre client on
77 the re-exporting nodes to be mounted with "32bitapi" mount option
79 --------------------------------------------------------------------------------
81 10-31-2013 Intel Corporation
83 * See https://wiki.hpdd.intel.com/display/PUB/Lustre+Support+Matrix
84 * Server builds for kernels:
85 2.6.32-358.18.1.el6 (RHEL6.4)
86 * Client builds for unpatched kernels:
87 2.6.32-358.18.1.el6 (RHEL6.4)
88 3.0.92-0.8 (SLES11 SP3)
90 * Recommended e2fsprogs version: 1.42.7.wc2 or newer
91 * NFS export disabled when stack size < 8192 (32-bit Lustre clients),
92 since the NFSv4 export of Lustre filesystem with 4K stack may cause a
93 stack overflow. For more information, please refer to bugzilla 17630.
94 * NFSv4 reexport to 32-bit NFS client nodes requires Lustre client on
95 the re-exporting nodes to be mounted with "32bitapi" mount option
97 --------------------------------------------------------------------------------
99 05-30-2013 Intel Corporation
101 * See https://wiki.hpdd.intel.com/display/PUB/Lustre+Support+Matrix
102 * Server builds for kernels:
103 2.6.32-358.6.2.el6 (RHEL6.4)
104 * Client builds for unpatched kernels:
105 2.6.32-358.6.2.el6 (RHEL6.4)
106 3.0.74-0.6.6 (SLES11 SP2)
108 * Recommended e2fsprogs version: 1.42.7.wc1 or newer
109 * NFS export disabled when stack size < 8192 (32-bit Lustre clients),
110 since the NFSv4 export of Lustre filesystem with 4K stack may cause a
111 stack overflow. For more information, please refer to bugzilla 17630.
112 * NFSv4 reexport to 32-bit NFS client nodes requires Lustre client on
113 the re-exporting nodes to be mounted with "32bitapi" mount option
115 --------------------------------------------------------------------------------
117 9-30-2012 Whamcloud, Inc.
119 * See https://wiki.hpdd.intel.com/display/PUB/Lustre+Support+Matrix
120 * Server support for kernels:
121 2.6.32-279.5.1.el6 (RHEL6)
122 * Client support for unpatched kernels:
123 2.6.32-279.5.1.el6 (RHEL6)
125 2.6.18-238.19.1.el5 (RHEL5)
126 2.6.16.60-0.69.1 (SLES10)
127 * Recommended e2fsprogs version: 1.41.90.wc4
128 * NFS export disabled when stack size < 8192 (32-bit Lustre clients),
129 since the NFSv4 export of Lustre filesystem with 4K stack may cause a
130 stack overflow. For more information, please refer to bugzilla 17630.
131 * NFSv4 reexport to 32-bit NFS client nodes requires Lustre client on
132 the re-exporting nodes to be mounted with "32bitapi" mount option
134 Severity : enhancement
136 Description: Update RHEL6.3 kernel to 2.6.32-279.5.1.el6.
138 Severity : enhancement
140 Description: Update RHEL6.3 kernel to 2.6.32-279.2.1.el6.
142 Severity : enhancement
144 Description: Update RHEL6.2 kernel to 2.6.32-220.17.1.el6
146 Severity : enhancement
148 Description: Update RHEL6.2 kernel to 2.6.32-220.13.1.el6
150 Severity : enhancement
152 Description: Update RHEL6.2 kernel to 2.6.32-220.7.1.el6
154 --------------------------------------------------------------------------------
156 04-30-2012 Whamcloud, Inc.
158 * See https://wiki.hpdd.intel.com/display/PUB/Lustre+Support+Matrix
159 * Server support for kernels:
160 2.6.32-220.4.2.el6 (RHEL6)
161 * Client support for unpatched kernels:
162 2.6.32-220.4.2.el6 (RHEL6)
164 2.6.18-238.19.1.el5 (RHEL5)
165 2.6.16.60-0.69.1 (SLES10)
166 * Recommended e2fsprogs version: 1.41.90.wc4
167 * NFS export disabled when stack size < 8192 (32-bit Lustre clients),
168 since the NFSv4 export of Lustre filesystem with 4K stack may cause a
169 stack overflow. For more information, please refer to bugzilla 17630.
170 * NFSv4 reexport to 32-bit NFS client nodes requires Lustre client on
171 the re-exporting nodes to be mounted with "32bitapi" mount option
173 Severity : enhancement
175 Description: Update RHEL6.2 kernel to 2.6.32-220.4.2.el6
177 Severity : enhancement
179 Description: Update RHEL6.2 kernel to 2.6.32-220.4.1.el6
181 Severity : enhancement
183 Description: Update RHEL6.2 kernel to 2.6.32-220.el6
185 Severity : enhancement
187 Description: Update RHEL6.1 kernel to 2.6.32-131.17.1.el6
189 --------------------------------------------------------------------------------
191 09-30-2011 Whamcloud, Inc.
193 * Server support for kernels:
194 2.6.18-238.19.1.el5 (RHEL5)
195 2.6.32-131.6.1.el6 (RHEL6)
196 * Client support for unpatched kernels:
197 (see http://wiki.lustre.org/index.php?title=Patchless_Client)
198 2.6.18-238.19.1.el5 (RHEL5)
199 2.6.32-131.6.1.el6 (RHEL6)
200 2.6.29.4-167.fc11 (FC11)
201 2.6.16.60-0.69.1 (SLES10)
203 * Recommended e2fsprogs version: 1.41.90.wc3
204 * NFS export disabled when stack size < 8192. Since the NFSv4 export of
205 Lustre filesystem with 4K stack may cause a stack overflow. For more
206 information, please refer to bugzilla 17630.
207 * NFSv4 reexport requires Lustre client on the re-exporting nodes to be
208 mounted with "32bitapi" mount option
210 Severity : enhancement
211 Description: Update RHEL6.1 kernel to 2.6.32-131.6.1.el6
213 Severity : enhancement
214 Description: Update RHEL5.6 kernel to 2.6.18-238.19.1.el5
216 Severity : enhancement
217 Description: Update RHEL5.6 kernel to 2.6.18-238.9.1.el5
220 Frequency : only with SLES10
222 Description: Use OFED "KMP" provided by Novell
223 Details : SLES10 SP3 ships with OFED in a separate "KMP" package.
224 Lustre is now built against this package. That means you need to
225 install the ofed-kmp package from Novell for the patchless client.
226 Note that the ofed-kmp that Novell ships may not exactly match the
227 kernel version but should still be compatible.
229 Severity : enhancement
231 Description: Update RHEL5.5 kernel to 2.6.18-194.17.1.el5
232 Update OEL5.5 kernel to 2.6.18-194.17.1.0.1.el5.
234 --------------------------------------------------------------------------------
236 2010-07-15 Oracle, Inc.
238 * Server support for kernels:
239 2.6.18-164.11.1.el5 (RHEL 5)
240 2.6.18-164.11.1.0.1.el5 (OEL 5)
241 * Client support for unpatched kernels:
242 (see http://wiki.lustre.org/index.php?title=Patchless_Client)
243 2.6.18-164.11.1.el5 (RHEL 5),
244 2.6.18-164.11.1.0.1.el5 (OEL 5)
245 2.6.16.60-0.42.8 (SLES 10),
247 2.6.29.4-167.fc11 (FC11)
248 * Recommended e2fsprogs version: 1.41.10-sun2
250 Severity : enhancement
252 Description: Distribute lfs_migrate script with lustre
253 Details : The lfs_migrate mentioned in the manual is now included in the
254 lustre tree in lustre/scripts and will distributed in lustre RPMs.
256 Severity : enhancement
258 Description: Failover nid list added to import proc file.
260 Severity : enhancement
262 Description: "som_preview" mount option required to use SOM
263 Details : The clients will need to use som_preview mount option
267 Description: Lustre does not do 1MB IOs to HW RAID
268 Details : Bump MAX_PHYS/HW_SEGMENTS and SG_ALL to 256 in the RHEL5 kernel.
270 Severity : enhancement
272 Description: lfs quota output is not very convenient for awk/sed-parsing
273 Details : Some positions in lfs quota output table could be empty or
274 non-empty which made it hard to parse it with scripts, now
275 a dash is put instead of space where there is not supposed
280 Description: Update to OFED 1.5.1
281 Details : OFED 1.5.1 will be used in environments that don't use
282 vendor supplied OFED.
286 Description: Remove lnet. prefix from lctl params display
287 Details : Also renamed lnet.memused to lnet_memused to
292 Description: Use conf_param -d to remove permanant settings.
296 Description: Add list_param -R
297 Details : lctl list_param now supports recursive mode.
300 Bugzilla : 21829, 22312
301 Description: Fix llobdstat.
302 Details : 1. Now updating the write_bytes counter
303 2. Port script fixes from 1.8.
307 Description: decrease the usage of memory on clients.
308 Details : 1. On clients, recycle dentries and inodes unused.
309 2. Delete the code related to ll_deathrow(att 6215 in bug 1443). It
314 Description: llobdstat fix and enhancement
315 Details : add a counter to set a limit to how many samples will be returned
316 fix a wildcard in the path to limit to obdfilter stats only
320 Description: lfs check only allowed for root
321 Details : Code cleanups and added a regresion test of non-root use of
326 Description: MDS Oopses in ptlrpc_dispatch_difficult_reply()/__list_add()
327 Details : Fix hr->hr_index overload.
331 Description: Fix FIEMAP functionality.
335 Description: FC11 patchless client support added
339 Description: Update RHEL5.4 kernel to 2.6.18-164.11.1.el5 and
340 OEL5.4 kernel to 2.6.18-164.11.1.0.1.el5.
344 Description: New acc-sm script and various test-framework enhancments.
348 Description: Previously landed patch was reverted.
352 Description: Update SLES10 kernel to 2.6.16.60-0.42.8 (patchless)
356 Description: Add ldiskfs maxdirsize mount option.
360 Description: Fix mkfs.lustre support for 16TB LUNs.
364 Description: Print mmp_check_interval
365 Details : print mmp_check_interval and make it possible to abort
366 mount operation in case it takes too long.
370 Description: Add cascading_rw to lustre mpi tests.
374 Description: Store SOM data in EA.
375 Details : Also several SOM related cleanups.
379 Description: Add refcount to md_open_data
383 Description: Use refcount on lov_request_set to prevent memory corruption.
387 Description: Statahead: use parent dir's i_mutex for synchronization.
391 Description: loadgen improvements.
395 Description: Enable peer health detection in ksocklnd and ko2iblnd.
399 Description: Control DCACHE_LUSTRE_INVALID flag with MDS_INODE_LOOKCUP lock
403 Description: Fix race on obd_devcie:md_stats
407 Description: Fix SLES 11 client builds.
408 Details : Resolved uninitalized pointer gcc warnings.
412 Description: Allow non-root access for "lfs check".
413 Details : Added a check in obd_class_ioctl() for OBD_IOC_PING_TARGET.
417 Description: Clear imp_force_reconnect correctly in ptlrpc_connect_interpret()
419 Severity : enhancement
421 Description: Update kernel to RHEL5.4 2.6.18-164.6.1.el5 and
422 OEL5 2.6.18-164.6.1.0.1.el5(Both in-kernel OFED enabled).
426 Description: Sanity test_27p fails.
427 Details : OBD_FAIL_OST_ENOSPC has been moved to filter_statfs().
428 Sanity tests 27 were cleaned up and improved.
432 Description: Unify req format on client/servers
433 Details : Use new req_capsule API [almost] everywhere instead of old PTLRPC
434 buffers and swabbers approach..
439 Description: drop join file stuff.
441 Severity : enhancement
443 Description: Add 2.6.27 and SLES11 (patchless client) support
447 Description: Fix in ptlrpc_expire_one_request() to print the signed time
450 Severity : enhancement
452 Description: Build kernels (RHEL5, OEL5 and SLES10/11) using the vendor's own
455 Severity : enhancement
457 Description: Remove set_info(KEY_UNLINKED) from MDS/OSC
459 Severity : enhancement
461 Description: correctly handle big reply message.
462 Details : send LNet event if reply is bigger then buffer and adjust this buffer
467 Description: Drop unnecessary __GFP_NOMEMALLOC flag from filter_get_page()
469 Severity : enhancement
471 Description: Add support for OEL5.
475 Description: Conf-sanity.sh 50g test - deactivated OST should not cause a panic.
477 Severity : enhancement
479 Description: "weak-modules" support
480 Details : Implement "weak-modules" support which enables kernel modules
481 to be used with any kernel that implements the same kABI. In
482 order to achieve this modules are now installed in
483 /lib/modules/$(uname -r)/updates/kernel on all distributions.
487 Description: client could not reconnect to OST because of an active request.
488 Details : The client now retries to reconnect to the same server, if a
489 connect request failed with EBUSY or -EAGAIN. Test case 35b
490 conf-sanity.sh has been created.
494 Description: Don't allow make backward step on assignin osc next id.
495 Details : race between allocation next id and ll_sync thread can be cause
496 of set wrong osc next id and can be kill valid ost objects.
500 Description: Changes in raid5-large-io-rhel5.patch to calculate sectors properly
504 Description: Increase the default BLK_DEF_MAX_SECTORS value for RHEL5 and SLES11
508 Description: Error handling in osc_statfs_interpret() has been improved.
509 Details : Check in osc_statfs_interpret() for EBADR.
513 Description: Do not send statfs() requests to OSTs disabled by administrator.
514 Details : Check in lov_prep_statfs_set() for non-NULL ltd_exp.
516 Severity : enhancement
518 Description: Added sanity-quota test to for lfs(1) to work with numeric uid/gid.
520 Severity : enhancement
522 Description: Update kernel to RHEL5 2.6.18-128.7.1.el5.
524 Severity : enhancement
526 Description: Enhancement for lfs(1) command to use numeric uid/gid.
530 Description: Improved cleanup procedure for conf-sanity test_46a.
534 Description: Do not start bulk transfer, if client tried to reconnect.
535 Details : Do not start bulk transfer, if client tried to reconnect,
536 but failed due to active rpcs. Check for exp_abort_active_req
537 has been implemented.
541 Description: Prevent inconsistences between linux and lustre mount structures.
542 Details : Wait indefinitely in server_wait_finished() until mnt_count drops.
543 Make the sleep interruptible.
545 Severity : enhancement
547 Description: provide server to client comms path
548 Details : server to client communications path via new LDLM_SET_INFO rpc
552 Description: truncate starts GFP_FS allocation under transaction causing deadlock
553 Details : ldiskfs_truncate calls grab_cache_page which may start page
554 allocation under an open transaction. This may lead to
555 calling prune_icache with consequent lustre reentrance.
559 Description: send correctly lsm on open replay
560 Details : MDS is trust to LSM size on replay open, but client can set wrong size
565 Description: enable client interface failover
566 Details : When a child reconnects from another NID, properly update export
567 nid hash position and ldlm reverse import.
569 Severity : enhancement
571 Description: Communicate OST degraded/readonly state via statfs to MDS
572 Details : Flags in the statfs returned from OSTs indicate whether the
573 OST is in a degraded RAID state, or if the filesystem has
574 turned read-only after a filesystem error is detected.
579 Description: don't panic if EPROTO was hit when reading symlink
580 Details : correctly handling request reference in error cases.
582 Severity : enhancement
584 Description: Add LustreNetLink, a kernel-userspace communcation path.
586 Severity : enhancement
588 Description: Update kernel to SLES10 SP2 2.6.16.60-0.39.3.
593 Description: access to llog context before init.
594 Details : move handling CATALOGS file at osc layer and forbid access to llog
598 Frequency : rare, only when client is writing to OST on local node
600 Description: Avoid deadlock for local client writes
601 Details : Use new OBD_BRW_MEMALLOC flag to notify OST about writes in the
602 memory freeing context. This allows OST threads to set the
603 PF_MEMALLOC flag on task structures in order to allocate memory
604 from reserved pools and complete IO.
605 Use GFP_HIGHUSER for OST allocations for non-local client writes,
606 so that the OST threads generate memory pressure and allow
607 inactive pages to be reclaimed.
609 Severity : enhancement
611 Description: Update kernel to RHEL5.3 2.6.18-128.1.14.el5.
616 Description: lock ordering violation between &cli->cl_sem and _lprocfs_lock
617 Details : move ldlm namespace creation in setup phase to avoid grab
618 _lprocfs_lock with cli_sem held.
622 Description: Temporarily disable grant shrink.
623 Details : Disable the feature for debugging.
626 Frequency : only when formatting test filesystems
628 Description: Unable to run several mkfs.lustre on loop devices at the same time
629 Details : mkfs.lustre returns error 256 on the concurrent loop devices
630 formatting. The solution is to proper handle the error.
632 Severity : enhancement
634 Description: Update kernel to RHEL5.3 2.6.18-128.1.6.el5.
636 Severity : enhancement
638 Description: Update kernel to SLES10 SP2 2.6.16.60-0.37.
643 Description: resolve race between obd_disconnect and class_disconnect_exports
644 Details : if obd_disconnect will be called to already disconnected export he
645 forget release one reference and osc module can't unloaded.
647 Severity : enhancement
649 Description: Allow tuning service thread via /proc
650 Details : For each service a new
651 /proc/fs/lustre/{service}/*/thread_{min,max,started} entry is
652 created that can be used to set min/max thread counts, and get the
653 current number of running threads.
658 Descriptoin: don't return error if have partially created objects for file.
659 Details : lov_update_create_set uses set->set_success as index for created
660 objects, so if some requests will be failed, they will have hole
661 at end of array and qos_shrink_lsm for allocate correct lsm.
663 Severity : enhancement
665 Description: Update OFED support to 1.4.1
670 Description: Reduce small size read RPC
671 Details : Set read-ahead limit for every file and only do read-ahead when
672 available read-ahead pages are bigger than 1M to avoid small size
675 Severity : enhancement
677 Description: move AT tunable parameters for more consistent usage
678 Details : add AT tunables under /proc/sys/lustre, add to conf_param parsing
680 Severity : enhancement
682 Description: add lazystatfs mount option to allow statfs(2) to skip down OSTs
683 Details : allow skip disconnected ost for send statfs request and hide error
689 Description: fix racy locking of mballoc block bitmaps causing BUG
690 Details : The locking of the mballoc buddy bitmap and the in-memory
691 block bitmap was using two different spin locks in some
692 cases. This made it possible to incorrectly access the
693 mballoc bitmap while another process was modifying it,
694 causing a sanity assertion to fail. While no on-disk corruption
695 was reported, there was some risk of this happening.
698 Frequency : rare, on llog test 6
700 Descriptoin: don't allow connect to already connected import
701 Details : allowing connect to already connected import is hide connecting problem.
706 Descriptoin: allow kill process which wait statahead result
707 Details : for some reasons 'ls' can stick in waiting result from statahead,
708 in this case need way for kill this process.
710 Severity : enhancement
712 Description: Add state history info file, enhance import info file
713 Details : Track import connection state changes in a new osc/mdc proc file;
714 add overview-type data to the osc/mdc import proc file.
716 Severity : enhancement
718 Description: MDS create should not wait for statfs RPC while holding DLM lock.
720 Severity : enhancement
722 Description: Update to RHEL5U3 kernel-2.6.18-128.1.1.el5.
727 Descriptoin: OST grant too much space to client even there are not enough space.
728 Details : Client will shrink its grant cache to OST if there are no write
729 activity over 6 mins (GRANT_SHRINK_INTERVAL), and OST will retrieve
730 this grant cache if there are already not enough avaible space
731 (left_space < total_clients * 32M).
734 Frequency : start MDS on uncleanly shutdowned MDS device
736 Descriptoin: ll_sync thread stay in waiting mds<>ost recovery finished
737 Details : stay in waiting mds<>ost recovery finished produce random bugs
738 due race between two ll_sync thread for one lov target. send
739 ACTIVATE event only if connect realy finished and import have
743 Frequency : rare, connect and disconnect target at same time
745 Descriptoin: ASSERTION(atomic_read(&imp->imp_inflight) == 0
746 Details : don't call obd_disconnect under lov_lock. this long time
747 operation and can block ptlrpcd which answer to connect request.
752 Descriptoin: don't lose wakeup for imp_recovery_waitq
753 Details : recover_import_no_retry or invalidate_import and import_close can
754 both sleep on imp_recovery_waitq, but we was send only one wakeup
758 Frequency : always with long access acl
760 Descriptoin: mds can't pack reply with long acl.
761 Details : mds don't control size of acl but they limited by reint/getattr
764 Severity : enhancement
766 Description: Update to SLES10 kernel-2.6.16.60-0.33.
768 Severity : enhancement
770 Description: Update to RHEL5 kernel-2.6.18-92.1.22.el5.
773 Frequency : start MDS on uncleanly shutdowned MDS device
775 Descriptoin: aborting recovery hang on MDS
776 Details : don't throttle destroy RPCs for the MDT.
779 Frequency : on remount
781 Description: external journal device not working after the remount
782 Details : clear dev_rdonly flag for external journal devices in
788 Description: shutdown vs evict race
789 Details : client_disconnect_export vs connect request race.
790 if client will evicted at this time - we start invalidate
791 thread without referece to import and import can be freed
795 Frequency : rare, need acl's on inode.
797 Description: client can't handle ost additional correctly
798 Details : if ost was added after client connected to mds client can have
799 hit lnet_try_match_md ... to big messages to wide striped files.
800 in this case need teach client to handle config events about add
801 lov target and update client max ea size at that event.
803 Severity : enhancement
805 Description: Changelogs
806 Details : Changelogs are a lightweight mechanism to track filesystem
807 metadata and namespace changes. The changelog is recorded
808 permanently on the MDTs, and is periodically "consumed" / purged
809 when records are no longer needed.
811 Severity : enhancement
813 Description: compact fld format with extents
814 Details : Store range of seq rather than every seq in FLD. Seq
815 controller update FLD rather than clients. In Case of CMD, mdt0
816 has FLD, all other metadata server act as non persistent proxy
817 for FLD queries and cache fld entries in fld cache.
822 Description: don't skip ost target if they assigned to file
823 Details : Drop slow OSCs if we can, but not for requested start idx.
824 This means "if OSC is slow and it is not the requested
825 start OST, then it can be skipped, otherwise skip it only
826 if it is inactive/recovering/out-of-space.
830 Description: more cleanup in mds_lov
831 Details : not send LOV EA under replay, we can't know about they size at this
832 time. Don't allow client connect to mds before any ost connected,
833 for avoid problems with LOV EA size and returning EIO to client.
835 Severity : enhancement
837 Description: Interoperability at server side (Disk interoperability)
839 Severity : enhancement
841 Description: Update to RHEL5 kernel-2.6.18-92.1.17.el5.
843 Severity : enhancement
845 Description: Update to SLES10 SP2 kernel-2.6.16.60-0.31.
847 Severity : enhancement
849 Description: New client IO stack (CLIO).
851 Severity : enhancement
853 Description: Commit on sharing. Eliminate inter-client dependencies between
854 uncommitted transactions by doing transaction commits.
855 Thereby clients may recovery independently.
858 Frequency : Create a symlink file with a very long name
860 Description: ldlm_cancel_pack()) ASSERTION(max >= dlm->lock_count + count)
861 Details : If there is no extra space in the request for early cancels,
862 ldlm_req_handles_avail() returns 0 instead of a negative value.
864 Severity : enhancement
866 Description: Add /proc entry for import status
867 Details : The mdc, osc, and mgc import directories now have
868 an import directory that contains useful import data for debugging
871 Severity : enhancement
873 Description: Re-disable certain /proc logging
874 Details : Enable and disable client's offset_stats, extents_stats and
875 extents_stats_per_process stats logging on the fly.
878 Frequency : Only on FC kernels 2.6.22+
880 Description: oops in statahead
881 Details : Do not drop reference count for the dentry from VFS when lookup,
882 VFS will do that by itself.
884 Severity : enhancement
886 Description: Generic /proc file permissions
887 Details : Set /Proc file permissions in a more generic way to enable non-
888 root users operate on some /proc files.
892 Description: Hitting mdc_commit_close() ASSERTION
893 Details : Properly handle request reference release in
894 ll_release_openhandle().
898 Description: quota recovery deadlock during mds failover
899 Details : This patch includes att18982, att18236, att18237 in bz14840.
901 1. fix osts hang when mds does failover with quotaon
902 2. prevent watchdog storm when osts threads wait for the
907 Frequency : only patchless client
908 Description: add workaround for race between add/remove dentry from hash
910 Severity : enhancement
912 Description: Allow OST glimpses to return PW locks
916 Description: LBUG when llog conf file is full
917 Details : When llog bitmap is full, ENOSPC should be returned for plain log.
921 Description: Prevent import from entering FULL state when server in recovery
925 Description: service mount cannot take device name with ":"
926 Details : Only when device name contains ":/" will mount treat it as
932 Description: replace ptlrpcd with the statahead thread to interpret the async
933 statahead RPC callback
937 Frequency : on recovery
938 Description: I/O failures after umount during fail back
939 Details : if client reconnected to restarted server we need join to recovery
940 instead of find server handler is changed and process self
941 eviction with cancel all locks.
943 Severity : enhancement
945 Description: Update to RHEL5 kernel-2.6.18-92.1.10.el5.
947 Severity : enhancement
949 Description: Update to SLES10 SP2 kernel-2.6.16.60-0.27.
951 Severity : enhancement
953 Description: Upcall on Lustre log has been dumped
954 Details : Allow for a user mode script to be called once a Lustre log has
955 been dumped. It passes the filename of the dumped log to the
956 script, the location of the script can be specified via
957 /proc/sys/lnet/debug_log_upcall.
962 Description: avoid idr_remove called for id which is not allocated.
963 Details : Move assigment s_dev for clustered nfs to end of initialization,
964 to avoid problem with error handling.
969 Description: avoid Already found the key in hash [CONN_UNUSED_HASH] messages
970 Details : When connection is reused this not moved from CONN_UNUSED_HASH
971 into CONN_USED_HASH and this prodice warning when put connection
972 again in unused hash.
975 Severity : enhancement
977 Description: File striping can now be set to use an arbitrary pool of OSTs.
979 Severity : enhancement
981 Description: Export bytes_read/bytes_write count on OSC/OST.
985 Description: Early reply size mismatch, MGC loses connection
986 Details : Apply the MGS_CONNECT_SUPPORTED mask at reconnect time so
987 the connect flags are properly negotiated.
992 Description: quotas are not honored with O_DIRECT
993 Details : all writes with the flag O_DIRECT will use grants which leads to
994 this problem. Now using OBD_BRW_SYNC to guard this.
998 Description: add quota statistics
999 Details : 1. sort out quota proc entries and proc code.
1000 2. add quota statistics
1002 Severity : enhancement
1004 Description: enable quota support for HEAD.
1008 Description: Properly propagate oinfo flags from lov to osc for statfs
1009 Details : restore missing copy oi_flags to lov requests.
1011 Severity : enhancement
1013 Description: Add man pages for llobdstat(8), llstat(8), plot-llstat(8),
1014 : l_getgroups(8), lst(8), routerstat(8)
1015 Details : included man pages for llobdstat(8), llstat(8),
1016 : plot-llstat(8), l_getgroups(8), lst(8), routerstat(8)
1018 Severity : enhancement
1020 Description: configure's --enable-quota should check the
1021 : kernel .config for CONFIG_QUOTA
1022 Details : configure is terminated if --enable-quota is passed but
1023 : no quota support is in kernel
1027 Description: Remove portals compatibility
1028 Details : Remove portals compatibility, not interoperable with releases
1033 Description: Resolve device initialization race
1034 Details : Prevent proc handler from accessing devices added to the
1035 obd_devs array but yet be intialized.
1037 Severity : enhancement
1039 Description: Update to SLES10 SP2 kernel-2.6.16.60-0.23.
1041 Severity : enhancement
1043 Description: Update to RHEL5 kernel-2.6.18-92.1.6.el5.
1048 Description: Using wrong pointer in osc_brw_prep_request
1049 Details : Access to array[-1] can produce panic if kernel compiled with
1050 CONFIG_PAGE_ALLOC enabled
1054 Description: Client runs out of low memory
1055 Details : Consider only lowmem when counting initial number of llap pages
1059 Description: *optional* service tags registration
1060 Details : if the "service tags" package is installed on a Lustre node
1061 When the filesystem is mounted, a local-node service tag will
1062 be created. See http://inventory.sun.com/ for more information
1063 about the Service Tags asset management system.
1067 Description: Kernel BUG tries to release flock
1068 Details : Lustre does not destroy flock lock before last reference goes
1069 away. So always drop flock locks when client is evicted and
1070 perform unlock regardless of successfulness of speaking to MDS.
1074 Description: add refcount protection for osc callbacks, avoid panic on shutdown
1078 Description: sanity test 65a fails if stripecount of -1 is set
1079 Details : handle -1 striping on filesystem in ll_dirstripe_verify
1084 Description: ASSERTION(CheckWriteback(page,cmd)) failed
1085 Details : badly clear PG_Writeback bit in ll_ap_completion can produce false
1088 Severity : enhancement
1090 Description: Update to RHEL5 kernel-2.6.18-53.1.21.el5.
1094 Description: do not process already freed flock
1095 Details : flock can possibly be freed by another thread before it reaches
1096 to ldlm_flock_completion_ast.
1100 Description: LBUG during stress test
1101 Details : Need properly lock accesses the flock deadlock detection list.
1105 Description: oops in page fault handler
1106 Details : kernel page fault handler can return two special 'pages' in error
1107 case, don't try dereference NOPAGE_SIGBUS and NOPAGE_OMM.
1111 Description: timeout with invalidate import.
1112 Details : ptlrpcd_check call obd_zombie_impexp_cull and wait request which
1113 should be handled by ptlrpcd. This produce long age waiting and
1114 -ETIMEOUT ptlrpc_invalidate_import and as result LASSERT.
1116 Severity : enhancement
1118 Description: Update to RHEL5 kernel-2.6.18-53.1.19.el5.
1122 Description: enable MGS and MDT services start separately
1123 Details : add a 'nomgs' option in mount.lustre to enable start a MDT with
1124 a co-located MGS without starting the MGS, which is a complement
1125 to 'nosvc' mount option.
1129 Frequency : after recovery
1130 Description: precreate to many object's after del orphan.
1131 Details : del orphan st in oscc last_id == next_id and this triger growing
1132 count of precreated objects. Set flag LOW to skip increase count
1133 of precreated objects.
1137 Frequency : rare, on clear nid stats
1138 Description: ASSERTION(client_stat->nid_exp_ref_count == 0)
1139 Details : when clean nid stats sometimes try destroy live entry,
1140 and this produce panic in free.
1144 Description: Stack overflow during MDS log replay
1145 ease stack pressure by using a thread dealing llog_process.
1149 Description: wait until IO finished before start new when do lock cancel.
1150 Details : VM protocol want old IO finished before start new, in this case
1151 need wait until PG_writeback is cleared until check dirty flag and
1152 call writepages in lock cancel callback.
1154 Severity : enhancement
1156 Description: using special macro for print time and cleanup in includes.
1160 Description: mds_mfd_close() ASSERTION(rc == 0)
1161 Details : In mds_mfd_close(), we need protect inode's writecount change
1162 within its orphan write semaphore to prevent possible races.
1166 Description: Obsolete CURRENT_SECONDS and use cfs_time_current_sec() instead.
1170 Frequency : rare, on shutdown ost
1171 Description: don't hit live lock with umount ost.
1172 Details : shrink_dcache_parent can be in long loop with destroy dentries,
1173 use shrink_dcache_sb instead.
1177 Description: don't panic with use echo client
1178 Details : echo client pass NULL as client nid pointer and this produce null
1179 pointer dereference.
1183 Description: fix build on ppc32
1184 Details : compile code with -m64 flag produce wrong object file for ppc32.
1188 Description: add message levels for liblustreapi
1192 Description: fix for occasional failure case of -ENOSPC in recovery-small tests
1193 Details : Move the 'good_osts' check before the 'total_bavail' check. This
1194 will result in an -EAGAIN and in the exit call path we call
1195 alloc_rr() which will with increasing aggressiveness attempt to
1196 aquire precreated objects on the minimum number of required OSCs.
1200 Description: Use old size assignment to avoid deadlock
1201 Details : Reverts the changes in bugs 2369 and bug 14138 that introduced
1202 the scheduling while holding a spinlock. We do not need locking
1203 for size in ll_update_inode() because size is only updated from
1204 the MDS for directories or files without objects, so there is no
1205 other place to do the update, and concurrent access to such inodes
1206 are protected by the inode lock.
1210 Description: resolve "_IOWR redefined" build error on SLES10
1214 Description: dump the memory debugging after all modules are unloaded to
1215 suppress false negative in conf_sanity test 39
1217 Severity : enhancement
1219 Description: build kernel-ib packages for OFED 1.3 in our release cycle
1224 Description: fix SLES kernel versioning
1225 Details : the kernel version for our SLES 10 kernel did not include a "-"
1226 before the "smp" at the end. while this was not a problem in
1227 general, it did mean that software trying to use the kernel
1228 version to try to detect a vendor specific kernel would fail.
1229 this was most evident by the OFED build scripts.
1233 Description: Don't update lov_desc members until making sure they are valid
1234 Details : When updating lov_desc members via proc fs, need fix their
1235 validities before doing the real update.
1239 Description: don't put request into delay list while invalidate in flight.
1240 Details : ptlrpc_delay_request sometimes put in delay list while invalidate
1241 import in flight. this produce timeout for invalidate and sometimes
1242 can cause stale data.
1246 Frequency : on ppc only
1247 Description: not convert ost objects for directory because it's not exist.
1248 Details : ll_dir_getstripe assume dirrectory has ost objects but this wrong.
1252 Description: Add FMODE_EXEC file flag for SLES10 SP1 kernel.
1254 Severity : enhancement
1256 Description: Update to support 2.6.22.14 vanilla kernel.
1260 Frequency : rare, on recovery
1261 Description: read procfs can produce deadlock in some situation
1262 Details : Holding lprocfs lock which send rpc can produce block for destroy
1263 obd objects and this also block reconnect with -EALREADY. This
1264 isn't fix all lprocfs bugs - but make it rare.
1266 Severity : enhancement
1268 Description: Update kernel to RHEL5 2.6.18-53.1.14.el5.
1271 Frequency : frequent on X2 node
1273 Description: mdc_set_open_replay_data LBUG
1274 Details : Set replay data for requests that are eligible for replay.
1278 Description: lustre_mgs: operation 101 on unconnected MGS
1279 Details : When MGC is disconnected from MGS long enough, MGS will evict the
1280 MGC, and late on MGC cannot successfully connect to MGS and a lot
1281 of the error messages complaining that MGS is not connected.
1285 Frequency : on network error
1286 Description: panic with double free request if network error
1287 Details : mdc_finish_enqueue is finish request if any network error ocuring,
1288 but it's true only for synchronus enqueue, for async enqueue
1289 (via ptlrpcd) this incorrect and ptlrpcd want finish request
1292 Severity : enhancement
1294 Description: client-side metadata stat-ahead during readdir(directory readahead)
1295 Details : perform client-side metadata stat-ahead when the client detects
1296 readdir and sequential stat of dir entries therein
1299 Frequency : on start mds
1301 Description: Implement get_info(last_id) in obdfilter.
1304 Frequency : occasional
1306 Description: Correctly check stale fid, not start epoch if ost not support SOM
1307 Details : open with flag O_CREATE need set old fid in op_fid3 because
1308 op_fid2 was overwritten with new generated fid, but mds can answer
1309 with one of these two fids and both is not stale. Setattr
1310 incorrectly started an epoch and assume will be called
1311 done_writing, but without SOM done_writing ever being called.
1314 Frequency : rare, depends on device drivers and load
1316 Description: MDS or OSS nodes crash due to stack overflow
1317 Details : Code changes in 1.8.0 increased the stack usage of some functions.
1318 In some cases, in conjunction with device drivers that use a lot
1319 of stack the MDS (or possibly OSS) service threads could overflow
1320 the stack. One change which was identified to consume additional
1321 stack has been reworked to avoid the extra stack usage.
1324 Frequency : occasional
1326 Description: Do not fail import if osc_interpret_create gets -EAGAIN
1327 Details : If osc_interpret_create got -EAGAIN it immediately exits and
1328 wakeup oscc_waitq. After wakeup oscc_wait_for_objects call
1329 oscc_has_objects and see OSC has no objests and call
1330 oscc_internal_create to resend create request.
1332 Severity : enhancement
1334 Description: Update to SLES10 SP1 latest kernel-2.6.16.54-0.2.5.
1336 Severity : enhancement
1338 Description: Update to RHEL5 latest kernel-2.6.18-53.1.13.el5.
1341 Frequency : very rare
1343 Description: Fix replay if there is an un-replied request and open
1344 Details : In some cases, older replay request will revert the
1345 mcd->mcd_last_xid on MDS which is used to record the client's
1346 latest sent request.
1348 Severity : enhancement
1350 Description: Update to RHEL5 latest kernel-2.6.18-53.1.6.el5.
1352 Severity : enhancement
1354 Description: Add rhel5 support to HEAD.
1356 Serverity : enhancement
1358 Description: Update RHEL4 kernel to 2.6.9-67.0.4.
1363 Description: Don't allow skipping OSTs if index has been specified.
1364 Details : Don't allow skipping OSTs if index has been specified, make
1365 locking in internal create lots better.
1369 Description: LBUG in ptlrpc_check_set() bad phase ebc0de00
1370 Details : access to bitfield in structure is always rounded to long
1371 and this produce problem with not atomic change any bit.
1375 Description: Lustre make rpms failed.
1376 Details : Remove ldiskfs spec file to avoids rpmbuild be confused when
1377 builds Lustre rpms from tarball.
1380 Frequency : rare on shutdown ost
1382 Description: If llog cancel was not send before clean_exports phase, this can
1383 produce deadlock in llog code.
1384 Details : If llog thread has last reference to obd and call class_import_put
1385 this produce deadlock because llog_cleanup_commit_master wait when
1386 last llog_commit_thread exited, but this never success because was
1387 called from llog_commit_thread.
1391 Description: allow userland application know is lost one of stripes.
1392 Details : fill lvb_blocks with error code on ost and return it to
1393 application if error flag found.
1397 Description: NULL lov_tgts causing MDS oops
1398 Details : more safe checks for NULL lov_tgts for avoid oops.
1400 Severity : enhancement
1402 Description: Update to RHEL4 latest kernel-2.6.9-67.0.1.EL.
1406 Descriptoin: make lov_create() will not stuck in obd_statfs_rqset()
1407 Details : If an OST is down the MDS will hang indefinitely in
1408 obd_statfs_rqset() waiting for the statfs data. While for
1409 MDS QOS usage of statfs, it should not stuck in waiting.
1411 Severity : enhancement
1413 Description: remote_acl support
1414 Details : Support ACL-based permission check for remote user.
1415 Support setfacl/getfacl for remote user with the utils
1416 "lfs {l,r}{s,g}etfacl" which follow the same parameter format as
1417 the system "{s,g}etfacl" utils.
1419 Severity : enhancement
1421 Description: Update to RHEL4 U6 kernel-2.6.9-67.EL.
1423 Severity : enhancement
1425 Description: Update to RHEL5 latest kernel-2.6.18-53.1.4.el5.
1429 Description: make mgs_setparam() handle fsname containing dash
1430 Details : fsname containing a dash does not work with lctl conf_param
1432 Severity : enhancement
1434 Description: Update to SLES10 SP1 latest kernel-2.6.16.54-0.2.3.
1436 Severity : enhancement
1438 Description: Update to RHEL5 Update-1 kernel 2.6.18-53.el5.
1442 Frequency : rare, at shutdown
1443 Description: access already free / zero obd_namespace.
1444 Details : if client_disconnect_export was called without force flag set,
1445 and exist connect request in flight, this can produce access to
1446 NULL pointer (or already free pointer) when connect_interpret
1447 store ocd flags in obd_namespace.
1451 Frequency : only at startup
1452 Description: not alloc memory with spinlock held.
1453 Details : allocation memory with GFP_KERNEL can produce sleep deadlock,
1454 if any spinlock held.
1456 Severity : enhancement
1458 Description: make lustre randomly fail allocating memory
1459 Details : Make lustre randomly failed allocating memory for testing purpose.
1461 Severity : enhancement
1463 Description: lost problems with lov objid file
1464 Details : Fixes some scability and access to not inited memory problems
1465 in work with lov objdid file.
1470 Description: lfs find does not continue on file error
1471 Details : Continue other files processing when a file/dir is absent.
1475 Description: Inconsistent usage of lustre_pack_reply()
1476 Details : Standardize the usage of lustre_pack_reply() such that it
1477 always generate a CERROR on failure.
1480 Frequency : occasional
1482 Description: MDS hang or stay in waiting lock
1483 Details : If client receive lock with CBPENDING flag ldlm need send lock
1484 cancel as separate rpc, to avoid situation when cancel request
1485 can't processed due all i/o threads stay in wait lock.
1489 Description: Update to RHEL5 kernel 2.6.18-8.1.15.el5.
1493 Description: Update to SLES10 SP1 kernel 2.6.16.53-0.16
1497 Description: Update to SLES9 kernel-2.6.5-7.287.3.
1501 Description: Update to RHEL4 latest kernel.
1503 Severity : enhancement
1505 Description: Build SLES10 patchless client fails
1506 Details : The configure was broken by run ./configure with
1507 --with-linux-obj=.... argument for patchless client. When the
1508 configure use --with-linux-obj, the LINUXINCLUDE= -Iinclude
1509 can't search header adequately. Use absolute path such as
1510 -I($LINUX)/include instead.
1514 Description: interrupt oig_wait produce painc on resend.
1515 Details : brw_redo_request can be used for resend requests from ptlrpcd and
1516 private set, and this produce situation when rq_ptlrpcd_data not
1517 copyed to new allocated request and triggered LBUG on assert
1518 req->rq_ptlrpcd_data != NULL. But this member used only for wakeup
1519 ptlrpcd set if request is changed and can be safety changed to use
1524 Description: LASSERT_{REQ,REP}SWAB macros are buggy
1525 Details : If SWAB_PARANOIA is disabled, the LASSERT_REQSWAB and
1526 LASSERT_REPSWAB macros become no-ops, which is incorrect. Drop
1527 these macros and replace them with their difinitions instead.
1531 Description: Update kernel patches for SLES10 2.6.16.53-0.8.
1532 Details : Update which_patch & target file for SLES10 latest kernel.
1535 Description: Remove client patches from SLES 10 kernel.
1536 Details : This causes SLES 10 clients to behave as patchless clients
1537 even on a Lustre-patched (server) kernel.
1539 Severity : enhancement
1541 Description: self-adjustable client's lru lists
1542 Details : use adaptive algorithm for managing client cached locks lru
1543 lists according to current server load, other client's work
1544 pattern, memory activities, etc. Both, server and client
1545 side namespaces provide number of proc tunables for controlling
1548 Severity : enhancement
1550 Description: light-weight GSS support
1551 Details : Support krb5n and krb5a mode, which keep Kerberos 5 authentication
1552 and reduce performance overhead.
1554 Severity : enhancement
1556 Description: Linux keyring support
1557 Details : Support using service of Linux keyring for Lustre GSS internal
1558 context refresh/cache mechanism.
1562 Description: Fix errors in lfs documentation
1563 Details : Fixes man pages
1567 Description: don't use GFP_* in generic Lustre code.
1568 Details : Use cfs_alloc_* functions and CFS_* flags for code portability.
1572 Description: obdclass is limited by single OBD_ALLOC(idarray)
1573 Details : replace OBD_ALLOC/OBD_FREE with OBD_VMALLOC/OBD_VFREE
1577 Description: warnings with build patchless client with vanila 2.6.19 and up
1578 Details : change old ctl_table style and replace ctl_table/ctl_table_header
1579 with cfs_sysctl_table_t/cfs_sysctl_table_header_t
1583 Frequency : Only for SLES
1584 Description: sanity_quota fail test_1
1585 Details : There are multiple occurences of $TSTUSR in SLES's /etc/group
1586 file, which makes TSTID[2] inunique.
1590 Frequency : Only for SLES9
1591 Description: Kernel patches for SLES9 2.6.5-7.286 kernel
1592 Details : Update target/ChangeLog/which_patch .
1596 Frequency : Only for test_10 in sanity-quota.sh
1597 Description: a bug in quota test code
1598 Details : checking a test flag in a wrong place
1602 Frequency : Only for test_13 in sanity-quota.sh
1603 Description: a bug in quota test script
1604 Details : shouldn't check overfull allocated quota
1608 Description: jbd statistics
1609 Details : Port older jbd statistics patch for sles10
1611 Severity : enhancement
1613 Description: Add printing inode info into message about error in writepage.
1618 Description: reply_lock_interpret crash due to race with it and lock cancel.
1619 Details : Do not replay locks that are being cancelled. Do not reference
1620 locks by their address during replay, just by their handle.
1624 Frequency : When flocks are used.
1625 Description: assertion failure in ldlm_cli_enquque_fini for non NULL lock.
1626 Details : Flock locks might destroy just granted lock if it could be merged
1627 with another existing flock, this is done in completion handler,
1628 so teach ldlm_cli_enquque_fini that this is a valid case for
1634 Description: Oops in read and write path when failing to allocate lock.
1635 Details : Check if lock allocation failed and return error back.
1639 Description: lstripe command fails for valid OST index
1640 Details : The stripe offset is compared to lov->desc.ld_tgt_count
1641 instead of lov->desc.ld_active_tgt_count.
1645 Description: sanity.sh failed test 103
1646 Details : RHEL mis-interpret setfacl "-X" param, so we won't test setfacl
1651 Description: df doesn't work properly if diskfs blocksize != 4K
1652 Details : Choose biggest blocksize of OST's as the LOV's blocksize.
1656 Description: lfs find on -1 stripe looping in lsm_lmm_verify_common()
1657 Details : Avoid lov_verify_lmm_common() on directory with -1 stripe count.
1659 Severity : enhancement
1661 Description: Adaptive timeouts
1662 Details : RPC timeouts adapt to changing server load and network
1663 conditions to reduce resend attempts and improve recovery time.
1667 Description: llapi_file_create() does not allow some changes
1668 Details : add llapi_file_open() that allows specifying the mode and
1669 open flags, and also returns an open file handle.
1673 Description: merge and cleanup kernel patches.
1674 Details : Remove mnt_lustre_list in vfs_intent-2.6-rhel4.patch.
1678 Description: Add journal checksum support.(Kernel part)
1679 Details : The journal checksum feature adds two new flags i.e
1680 JBD2_FEATURE_INCOMPAT_ASYNC_COMMIT and
1681 JBD2_FEATURE_COMPAT_CHECKSUM. JBD2_FEATURE_CHECKSUM flag
1682 indicates that the commit block contains the checksum for
1683 the blocks described by the descriptor blocks. Now commit
1684 record can be sent to disk without waiting for descriptor
1685 blocks to be written to disk. This behavior is controlled
1686 using JBD2_FEATURE_ASYNC_COMMIT flag.
1690 Description: OSS needs mutliple precreate threads
1691 Details : Add ability to start more than one create thread per OSS.
1695 Description: Sanity.sh test_65k failed: setstripe should have succeeded
1696 Details : execute lfs setstripe on client
1700 Description: mds_obd_create error creating tmp object
1701 Details : When the user sets quota on root, llog will be affected and can't
1702 create files and write files.
1705 Frequency : Always on ia64 patchless client, and possibly others.
1707 Description: Add EXPORT_SYMBOL check for node_to_cpumask symbol.
1708 Details : This allows the patchless client to be loaded on architectures
1709 without this export.
1713 Description: RedHat Update kernel for RHEL5
1714 Details : Add the kernel config file for RHEL5.
1718 Description: RedHat Update kernel for RHEL5
1719 Details : Modify the kernel config file more closer RHEL5.
1723 Description: Build failure against Centos5 (RHEL5)
1724 Details : Define PAGE_SIZE when it isn't present.
1728 Description: test_7 on sanity-quota.sh can't be used on separate machines
1729 Details : add facet to handle it
1733 Description: "ll_intent_file_open()) lock enqueue: err: -13" with nfs
1734 Details : with NFS, the anon dentry's parent was set to itself in
1735 d_alloc_anon(), so in MDS, we use rec->ur_fid1 to find the
1736 corresponding dentry other than use rec->ur_name.
1738 Severity : enhancement
1740 Description: lfs setstripe enhancement
1741 Details : Make lfs setstripe understand 'k', 'm' and 'g' for stripe size.
1745 Description: enable data checksumming by default
1746 Details : enable checksum by default, allow --disable-checksum
1747 configure option and "-o nochecksum" mount option. Checksums
1748 can also be disabled at runtime via $LPROC/osc/*/checksum_pages.
1752 Description: lustre support for RHEL5
1753 Details : Add support for RHEL5.
1757 Description: Client eviction due to failover config
1758 Details : after a connection loss, the lustre client should attempt to
1759 reconnect to the last active server first before trying the
1760 other potential connections.
1764 Description: when mds and osts use different quota unit(32bit and 64bit),
1765 quota will be released repeatly.
1766 Details : void sending multiple quota reqs to mds, which will keep the
1767 status between the reqs.
1771 Description: osts not allocated evenly to files
1772 Details : change the condition to increase offset_idx
1775 Frequency : Always for filesystems larger than 2TB on 32-bit systems.
1777 Description: Data corruption for OSTs that are formatted larger than 2TB
1779 Details : When generating the bio request for lustre file writes the
1780 sector number would overflow a temporary variable before being
1781 used for the IO. The data reads correctly from Lustre (which
1782 will overflow in a similar manner) but other file data or
1783 filesystem metadata may be corrupted in some cases.
1787 Description: Tune the kernel for good SCSI performance.
1788 Details : Set the value of /sys/block/{dev}/queue/max_sectors_kb
1789 to the value of /sys/block/{dev}/queue/max_hw_sectors_kb
1794 Description: rewrite ext2-derived code in obdclass/uuid.c
1795 Details : rewrite inherited code (uuid parsing code from ext2 utils)
1796 from scratch preserving functionality.
1800 Description: "lfs find -obd UUID" prints directories
1801 Details : "lfs find -obd UUID" will return all directory names instead
1802 of just file names. It is incorrect because the directories
1803 do not reside on the OSTs.
1807 Description: use i_size_read and i_size_write in 2.6 port
1808 Details : replace inode->i_size access with i_size_read/write()
1811 Frequency : when removing large files
1813 Description: scheduling issue during removal of large Lustre files
1814 Details : Don't take the BKL in fsfilt_ext3_setattr() for 2.6 kernels.
1815 It causes scheduling issues when removing large files (17TB in the
1819 Frequency : only with liblustre clients on XT3
1821 Description: evictions taking too long
1822 Details : allow llrd to evict clients directly on OSTs
1825 Frequency : only on ppc
1827 Description: /proc/fs/lustre/devices broken on ppc
1828 Details : The patch as applied to 1.6.2 doesn't look correct for all arches.
1829 We should make sure the type of 'index' is loff_t and then cast
1830 explicitly as needed below. Do not assign an explicitly cast
1834 Frequency : only for rhel5
1836 Description: Kernel patches update for RHEL5 2.6.18-8.1.10.el5.
1837 Details : Modify the target file & which_kernel.
1839 Severity : enhancement
1841 Description: omit set fsid for export NFS
1842 Details : fix set/restore device id for avoid EMFILE error and mark lustre
1843 fs as FS_REQUIRES_DEV for avoid problems with generate fsid.
1847 Frequency : Always, for kernels after 2.6.16
1848 Description: Fix warning idr_remove called for id=.. which is not allocated.
1849 Details : Last kernels save old s_dev before kill super and not allow
1850 to restore from callback - restore it before call kill_anon_super.
1854 Description: buffer overruns could theoretically occur
1855 Details : llapi_semantic_traverse() modifies the "path" argument by
1856 appending values to the end of the origin string, and a
1857 overrun may occur. Adding buffer overrun check in liblustreapi.
1861 Description: Fix error on 'ls .' at the top of the Lustre mount.
1862 Details : Don't revalidate dentry if it is a root dentry.
1866 Description: Kernel patches update for RHEL4 2.6.9-55.0.6.
1867 Details : Modify vm-tunables-rhel4.patch.
1871 Description: Kernel config for 2.6.18-vanilla.
1872 Details : Modify targets/2.6-vanilla.target.in.
1873 Add config file kernel-2.6.18-2.6-vanilla-i686.config.
1874 Add config file kernel-2.6.18-2.6-vanilla-i686-smp.config.
1875 Add config file kernel-2.6.18-2.6-vanilla-x86_64.config.
1876 Add config file kernel-2.6.18-2.6-vanilla-x86_64-smp.config.
1880 Description: improve handling recoverable errors
1881 Details : if request processig with error which can be recoverable on server
1882 request should be resend, otherwise page released from cache and
1887 Description: Kernel patches update for RHEL5 2.6.18-8.1.14.el5.
1888 Details : Modify target file & which_patch.
1889 A flaw was found in the IA32 system call emulation provided
1890 on AMD64 and Intel 64 platforms. An improperly validated 64-bit
1891 value could be stored in the %RAX register, which could trigger an
1892 out-of-bounds system call table access. An untrusted local user
1893 could exploit this flaw to run code in the kernel
1894 (ie a root privilege escalation). (CVE-2007-4573).
1898 Description: O_DIRECT bypasses client statistics.
1899 Details : When running with O_DIRECT I/O, neither the client rpc_stats nor
1900 read_ahead_stats were updated. copied the stats section from
1901 osc_send_oap_rpc() into async_internal().
1905 Description: Add jbd statistics patch for RHEL5 and 2.6.18-vanilla
1909 Description: change order of libsysio includes
1910 Details : '#include sysio.h' should always come before '#include xtio.h'
1914 Description: handle "serious error: objid * already exists" more gracefully
1915 Details : If LAST_ID value on disk is smaller than the objects existing in
1916 the O/0/d* directories, it indicates disk corruption and causes an
1917 LBUG(). If the object is 0-length, then we should use the existing
1918 object. This will help to avoid a full fsck in most cases.
1920 Severity : enhancement
1922 Description: adapt the lustre_config script to support the upgrade case
1923 Details : Add "-u" option for lustre_config script to support upgrading 1.4
1924 server targets to 1.6 in parallel.
1928 Description: To avoid grant space > avaible space when the disk is almost
1929 full. Without this patch you might see the error "grant XXXX >
1930 available" or some LBUG about grant, when the disk is almost
1932 Details : In filter_check_grant, for non_grant cache write, we should
1933 check the left space by if (*left > ungranted + bytes), instead
1934 of (*left > ungranted), because only we are sure the left space
1935 is enough for another "bytes", then the ungrant space should be
1936 increase. In client, we should update cl_avail_grant only there
1937 is OBD_MD_FLGRANT in the reply.
1941 Description: Update RHEL 4 kernel to fix local root privilege escalation.
1942 Details : Update to the latest RHEL 4 kernel to fix the vulnerability
1943 described in CVE-2007-4573. This problem could allow untrusted
1944 local users to gain root access.
1947 Frequency : when using O_DIRECT and quotas
1949 Description: Incorrect file ownership on O_DIRECT output files
1950 Details : block usage reported by 'lfs quota' does not take into account
1951 files that have been written with O_DIRECT.
1956 Description: touch file failed when fs is not full
1957 Details : OST in recovery should not be discarded by MDS in alloc_qos(),
1958 otherwise we can get ENOSP while fs is not full.
1962 Description: parallel lock callbacks
1963 Details : Instead of sending blocking and completion callbacks as separated
1964 requests, adding them to a set and sending in parallel.
1967 Frequency : only for Cray XT3
1968 Bugzilla : 12829/13455
1969 Description: Changing primary group doesn't change the group lustre assigns to
1971 Details : When CRAY_XT3 is defined, the fsgid supplied by the client is
1972 overridden with the primary group provided by the group upcall,
1973 whereas the supplied fsgid can be trusted if it is in the list of
1974 supplementary groups returned by the group upcall.
1976 Severity : enhancement
1978 Description: Allow masking D_WARNING, D_ERROR messages from console
1979 Details : Console messages can now be disabled via lnet.printk.
1983 Description: User code with malformed file open parameter crashes client node
1984 Details : Before packing join_file req, all the related reference should be
1985 checked carefully in case some malformed flags cause fake
1986 join_file req on client.
1990 Description: LDLM_ENQUEUE races with LDLM_CP_CALLBACK
1991 Details : ldlm_completion_ast() assumes that a lock is granted when the req
1992 mode is equal to the granted mode. However, it should also check
1993 that LDLM_FL_CP_REQD is not set.
1997 Description: Heavy nfs access might result in deadlocks
1998 Details : After ELC code landed, it is now improper to enqueue any mds
1999 locks under och_sem, because enqueue might want to decide to
2000 cancel open locks for same inode we are holding och_sem for.
2004 Description: Client eviction while running blogbench
2005 Details : A lot of unlink operations with concurrent I/O can lead to a
2006 deadlock causing evictions. To address the problem, the number of
2007 oustanding OST_DESTROY requests is now throttled to
2008 max_rpcs_in_flight per OSC and LDLM_FL_DISCARD_DATA blocking
2009 callbacks are processed in priority.
2013 Description: enable ACLs on MDS by default
2014 Details : ACLs must be enabled on MDS by default.
2017 Frequency : PPC/PPC64 only
2019 Description: conflicts between asm-ppc64/types.h and lustre_types.h
2020 Details : fix duplicated definitions between asm-ppc64/types.h and
2021 lustre_types.h on PPC.
2024 Frequency : PPC/PPC64 only
2026 Description: asm-ppc/segment.h does not exist
2027 Details : fix compile issue on PPC.
2031 Description: better handle error messages in extents code
2034 Frequency : RHEL4 only
2036 Description: mkfs is very slow on IA64/RHEL4
2037 Details : A performance regression has been discovered in the MPT Fusion
2038 driver between versions 3.02.73rh and 3.02.99.00rh. As a
2039 consequence, we have downgraded the MPT Fusion driver in the RHEL4
2040 kernel from 3.02.99.00 to 3.02.73 until this problem is fixed.
2042 Severity : enhancement
2044 Description: SNMP support enhancement
2045 Details : Adding total number of sampled request for an MDS node in snmp
2048 Severity : enhancement
2050 Description: Optimize ldlm waiting list processing for PR extent locks
2051 Details : When processing waiting list for read extent lock and meeting read
2052 lock that is same or wider to it that is not contended, skip
2053 processing rest of the list and immediatelly return current
2054 status of conflictness, since we are guaranteed there are no
2055 conflicting locks in the rest of the list.
2059 Description: Time out and refuse to reconnect
2060 Details : When the failover node is the primary node, it is possible
2061 to have two identical connections in imp_conn_list. We must
2062 compare not conn's pointers but NIDs, otherwise we can defeat
2063 connection throttling.
2067 Description: port llog fixes from b1_6 into HEAD
2068 Details : Port llog reference couting and some llog cleanups from b1_6
2069 (bug 10800) into HEAD, for protect from panic and access to already
2070 free llog structures.
2074 Description: Detect stride IO mode in read-ahead
2075 Details : When a client does stride read, read-ahead should detect that and
2076 read-ahead pages according to the detected stride pattern.
2080 Description: data checksumming impacts single node performance
2081 Details : add support for several checksum algorithm. Currently, only CRC32
2082 and Adler-32 are supported. The checksum type can be changed on
2083 the fly via /proc/fs/lustre/osc/*/checksum_type.
2087 Description: use adler32 for page checksums
2088 Details : when available, use the Adler-32 algorithm instead of CRC32 for
2093 Description: build for x2 fails
2094 Details : fix compile issue on Cray systems.
2098 Description: Properly match for duplicate locks
2099 Details : Due to different lock order from skiplists code, we need to
2100 traverse entire list for now
2103 Frequency : only on PPC/SLES10
2105 Description: "BITS_PER_LONG is not 32 or 64" in linux/idr.h
2106 Details : On SLES10/PPC, fs.h includes idr.h which requires BITS_PER_LONG to
2107 be defined. Add a hack in mkfs_lustre.c to work around this compile
2112 Description: LASSERT on MDS when client holding flock lock dies
2113 Details : ldlm pool logic depends on number of granted locks equal to
2114 number of released locks which is not true for flock locks, so
2115 just exclude such locks from consideration.
2119 Description: MDS deadlock with many ll_sync_lov threads and I/O stalled
2120 Details : Use fsfilt_sync() for both the whole filesystem sync and
2121 individual file sync to eliminate dangerous inode locking
2122 with I_LOCK that can lead to a deadlock.
2126 Description: performance in 1.6.3
2127 Details : Force q->max_phys_segments to MAX_PHYS_SEGMENTS on SLES10 to be
2128 sure that 1MB requests are not fragmented by the block layer.
2130 Severity : enhancement
2132 Description: organize the server-side client stats on per-nid basis
2133 Details : Change the structure of stats under obdfilter and mds to
2141 The "uuid"s file would list the uuids of _active_ exports.
2142 And the clear entry is to clear all stats and stale nids.
2144 Severity : enhancement
2146 Description: eliminate client locks in face of contention
2147 Details : file contention detection and lockless i/o implementation
2148 for contended files.
2152 Description: Reinitialize optind to 0 so that interactive lfs works in all cases
2155 Frequency : very rare, if additional xattrs are used on kernels >= 2.6.12
2157 Description: MDS may lose file striping (and hence file data) in some cases
2158 Details : If there are additional extended attributes stored on the MDS,
2159 in particular ACLs, SELinux, or user attributes (if user_xattr
2160 is specified for the client mount options) then there is a risk
2161 of attribute loss. Additionally, the Lustre file striping
2162 needs to be larger than default (e.g. striped over all OSTs),
2163 and an additional attribute must be stored initially in the
2164 inode and then increase in size enough to be moved to the
2165 external attribute block (e.g. ACL growing in size) for file
2170 Description: skiplist implementation simplification
2171 Details : skiplists are used to group compatible locks on granted list
2172 that was implemented as tracking first and last lock of each
2173 lock group the patch changes that to using doubly linked lists
2177 Description: MDS LBUG: ASSERTION(!IS_ERR(dchild))
2178 Details : Change LASSERTs to client eviction (i.e. abort client's recovery)
2179 because LASSERT on both the data supplied by a client, and the
2180 data on disk is dangerous and incorrect.
2182 Severity : enhancement
2184 Description: Slow truncate/writes to huge files at high offsets.
2185 Details : Directly associate cached pages to lock that protect those pages,
2186 this allows us to quickly find what pages to write and remove
2187 once lock callback is received.
2191 Description: more ldlm soft lockups
2192 Details : In ldlm_resource_add_lock(), call to ldlm_resource_dump()
2193 starve other threads from the resource lock for a long time in
2194 case of long waiting queue, so change the debug level from
2195 D_OTHER to the less frequently used D_INFO.
2197 Severity : enhancement
2199 Description: add -gid, -group, -uid, -user options to lfs find
2203 Description: Hung threads in invalidate_inode_pages2_range
2204 Details : The direct IO path doesn't call check_rpcs to submit a new RPC
2205 once one is completed. As a result, some RPCs are stuck in the
2206 queue and are never sent.
2210 Description: filter threads hungs on waiting journal commit
2211 Details : Cleanup filter group llog code, then only filter group llog will
2212 be only created in the MDS/OST syncing process.
2216 Description: Procfs and llog threads access destoryed import sometimes.
2217 Details : Sync the import destoryed process with procfs and llog threads by
2218 the import refcount and semaphore.
2220 Severity : enhancement
2222 Description: openlock cache of b1_6 port to HEAD
2227 Description: kernel BUG at ldiskfs2_ext_new_extent_cb
2228 Details : If insertion of an extent fails, then discard the inode
2229 preallocation and free data blocks else it can lead to duplicate
2234 Description: don't always update ctime in ext3_xattr_set_handle()
2235 Details : Current xattr code updates inode ctime in ext3_xattr_set_handle.
2236 In some cases the ctime should not be updated, for example for
2237 2.0->1.8 compatibility it is necessary to delete an xattr and it
2238 should not update the ctime.
2242 Bugzilla : 15713/16362
2243 Description: Assertion in iopen_connect_dentry in 1.6.3
2244 Details : looking up an inode via iopen with the wrong generation number can
2245 populate the dcache with a disconneced dentry while the inode
2246 number is in the process of being reallocated. This causes an
2247 assertion failure in iopen since the inode's dentry list contains
2248 both a connected and disconnected dentry.
2252 Description: assertion failure in ldlm_handle2lock()
2253 Details : fix a race between class_handle_unhash() and class_handle2object()
2254 introduced in lustre 1.6.5 by bug 13622.
2259 Description: Kernel BUG: sd_iostats_bump: unexpected disk index
2260 Details : remove the limit of 256 scsi disks in the sd_iostat patch
2265 Description: oops in sd_iostats_seq_show()
2266 Details : unloading/reloading the scsi low level driver triggers a kernel
2267 bug when trying to access the sd iostat file.
2272 Description: Kernel panics during QLogic driver reload
2273 Details : REQ_BLOCK_PC requests are not handled properly in the sd iostat
2274 patch, causing memory corruption.
2279 Description: journal_dev option does not work in b1_6
2280 Details : pass mount option during pre-mount.
2282 Severity : enhancement
2284 Description: Add a FIEMAP(FIle Extent MAP) ioctl
2285 Details : FIEMAP ioctl will allow an application to efficiently fetch the
2286 extent information of a file. It can be used to map logical blocks
2287 in a file to physical blocks in the block device.
2291 Description: LDLM soft lockups - improvement
2292 Details : It is be possible to send the lock handle along with each read
2293 or write request because the client is already doing a lock match
2294 itself so there isn't any reason the OST should have to re-do that
2298 Frequency : only on Cray X2
2300 Description: X2 build failures
2301 Details : fix build failures on Cray X2.
2305 Description: xid & resent requests
2306 Details : Initialize RPC XID from clock at startup (randomly if clock is
2309 Severity : enhancement
2311 Description: Add lustre_start utility to start or stop multiple Lustre servers
2316 Description: Lustre GPF in {:ptlrpc:ptlrpc_server_free_request+373}
2317 Details : In case of memory pressure, list_del() can be called twice on
2318 req->rq_history_list, causing a kernel oops.
2322 Description: kptllnd_peer_check_sends()) ASSERTION(!in_interrupt()) failed
2323 Details : fix stack overflow in the distributed lock manager by defering
2324 export eviction after a failed AST to the elt thread instead of
2325 handling it in the dlm interpret routine.
2329 Description: Convert some comments to new format.
2330 Details : Update documenting comments to match doxygen conventions.
2334 Description: Grammar fixes.
2335 Details : A couple of trivial spelling fixes.
2339 Description: OSD_COUNTERS-mandatory
2340 Details : Make previously optional ->oti_{w,r}_locks sanity checks mandatory
2341 to simplify the code.
2345 Description: simplify cmm_device freeing logic.
2346 Details : Call cmm_device_free() in the failure path of cmm_device_alloc().
2350 Description: Add lockdep support to dt_object_operations locking interface.
2351 Details : Augment ->do_{read,write}_lock() prototypes with a `role'
2352 parameter indicating lock ordering. Update mdd code to use new
2357 Description: Introduce failloc constants for lockless IO tests.
2358 Details : Add two new failloc constants to test lockless IO. Only one of
2359 them in implemented---another is checked in yet to be landed
2364 Description: Add lockdep support for inode mutex.
2365 Details : Introduce and use new LOCK_INODE_MUTEX_PARENT() macro to be used
2366 in the situations where Lustre has to lock more than one inode
2371 Description: Add optional invariants checking support.
2372 Details : Add new LINVRNT() macro, optional on new --enable-invariants
2373 configure switch. This macro is to be used for consistency and
2374 sanity checks that are too expensive to be left in `production'
2379 Description: Zap lock->l_granted_mode with explicit LCK_MINMODE.
2380 Details : Use LCK_MINMODE rather than 0 to reset lock->l_granted_mode to
2385 Description: Add lockdep support for ldlm_lock and ldlm_resource.
2386 Details : Use spin_lock_nested() in (the only) situation where more than
2387 one ldlm_lock is locked simultaneously. Also, fix possible
2388 dead-lock in ldlm_lock_change_resource() by enforcing particular
2393 Description: Use struct ldlm_callback_suite in ldlm_lock_create().
2394 Details : Instead of specifying each ldlm_lock call-back through separate
2395 parameter, wrap them into struct ldlm_callback_suite.
2399 Description: Kill join_lru obd method and its callers.
2400 Details : CLIO uses lock weighting policy to keep locks over mmapped regions
2401 in memory---a requirement implemented through ->o_join_lru() obd
2402 method in HEAD. Remove this method and its users.
2406 Description: Add asynchronous ldlm ENQUEUE completion handler.
2407 Details : CLIO posts ENQUEUE requests asynchronously through ptlrpcd---a
2408 case that stock ldlm_completion_ast() cannot handle as it waits
2409 until lock is granted. Introduce new ldlm_completion_ast_async()
2410 for this. Also comment ldlm_completion_ast().
2414 Description: ldlm_error <-> errno conversion.
2415 Details : Add functions to map (rather arbitrary) between LDLM error codes
2416 and standard errno values. CLIO needs this to prevent LDLM specific
2417 constants from escaping ldlm and osc.
2421 Description: Kill unused ldlm_handle2lock_ns() function.
2422 Details : Kill unused ldlm_handle2lock_ns() function.
2426 Description: Add lu_ref support to ldlm_lock
2427 Details : lu_ref support for ldlm_lock and ldlm_resource. See lu_ref patch.
2428 lu_ref fields ->l_reference and ->lr_reference are added to ldlm_lock
2429 and ldlm_resource. LDLM interface has to be changed, because code that
2430 releases a reference on a lock, has to "know" what reference this is.
2431 In the most frequent case
2433 lock = ldlm_handle2lock(handle);
2435 LDLM_LOCK_PUT(lock);
2437 no changes are required. When any other reference (received _not_ from
2438 ldlm_handle2lock()) is released, LDLM_LOCK_RELEASE() has to be called
2439 instead of LDLM_LOCK_PUT().
2441 Arguably, changes are pervasive, and interface requires some discipline
2442 for proper use. On the other hand, it was very instrumental in finding
2443 a few leaked lock references.
2447 Description: Add ldlm_lock_addref_try().
2448 Details : Introduce ldlm_lock_addref_try() function (used by CLIO) that
2449 attempts to addref a lock that might be being canceled
2454 Description: Add ldlm_weigh_callback().
2455 Details : Add new ->l_weigh_ast() call-back to ldlm_lock. It is called
2456 by ldlm_cancel_shrink_policy() to estimate lock "value", instead of
2457 hard-coded `number of pages' logic.
2461 Description: Add lockdep annotations to llog code.
2462 Details : Use appropriately tagged _nested() locking calls in the places
2463 where llog takes more than one ->lgh_lock lock.
2467 Description: Add loi_kms_set().
2468 Details : Wrap kms updates into a helper function.
2472 Description: Constify instances of struct lsm_operations.
2473 Details : Constify instances of struct lsm_operations.
2477 Description: lu_conf support.
2478 Details : On a server, a file system object is uniquely identified
2479 by a fid, which is sufficient to locate and load all object
2480 state (inode). On a client, on the other hand, more data are
2481 necessary instantiate an object. Change lu_object_find() and
2482 friends to take additional `lu_conf' argument describing object.
2483 Typically this includes layout information.
2487 Description: lu_context fixes.
2488 Details : Introduce new lu_context functions that are needed on the client
2489 side, where some system threads (ptlrpcd) are shared by multiple
2490 modules, and so cannot be stopped during module shutdown.
2494 Description: Add start and stop methods to lu_device_type_operations.
2495 Details : Introduce two new methods in lu_device_type_operations, that are
2496 invoked when first instance of a given type is created and last one
2497 is destroyed respectively. This is need by CLIO.
2501 Description: Add lu_ref support to struct lu_device.
2502 Details : Add lu_ref support to lu_object and lu_device. lu_ref is used to
2503 track leaked references.
2507 Description: Introduce lu_kmem_descr.
2508 Details : lu_kmem_descr and its companion interface allow to create
2509 and destroy a number of kmem caches at once.
2513 Description: Fix lu_object finalization race.
2514 Details : Fix a race between lu_object_find() finding an object and its
2515 concurrent finalization. This race is (most likely) not possible
2516 on the server, but might happen on the client.
2520 Description: Introduce lu_ref interface.
2521 Details : lu_ref is a debugging module allowing to track references to
2522 a given object. It is quite cpu expensive, and has to be
2523 explicitly enabled with --enable-lu_ref. See usage description
2528 Description: Factor lu_site procfs stats into a separate function.
2529 Details : Separate lu_site stats printing code into a separate function
2530 that can be reused on a client.
2534 Description: Constify instances of struct {lu,dt,md}_device_operations.
2535 Details : Constify instances of struct {lu,dt,md}_device_operations.
2539 Description: Introduce struct md_site and move meta-data specific parts of
2540 struct lu_site here.
2541 Details : Move md-specific fields out of struct lu_site into special struct
2542 md_site, so that lu_site can be used on a client.
2546 Description: Kill mdd_lov_destroy().
2547 Details : Remove unused mdd code.
2551 Description: Add st_block checking to multistat.c.
2552 Details : Add st_block checking to multistat.c.
2556 Description: Add lu_ref support to struct obd_device.
2557 Details : Add lu_ref tracking to obd_device.
2561 Description: Kill obd_set_fail_loc().
2562 Details : Remove unused code.
2566 Description: Add special type for ptlrpc_request interpret functions.
2567 Details : Add lu_env parameter to ->rq_interpreter call-back. NULL is passed
2568 there. Actual usage will be in CLIO.
2572 Description: Replace RW_LOCK_UNLOCKED() macro with rwlock_init().
2573 Details : Replace RW_LOCK_UNLOCKED() with rwlock_init() as the former
2574 doesn't work with lockdep.
2578 Description: Add rwv.c test program.
2579 Details : New testing program exercising readv(2) and writev(2) (Qian).
2583 Description: Add sendfile.c test program.
2584 Details : New testing program exercising sendfile(2) (Jay).
2588 Description: Ratelimit a message that can be very frequent.
2589 Details : Ratelimit a memory allocation failure message that can
2594 Description: Use cdebug_show() in CDEBUG-style macros defined outside of libcfs.
2595 Details : Use cdebug_show() in CDEBUG-style macros defined outside of libcfs.
2599 Description: Liblustre build fixes.
2600 Details : Liblustre build fixes.
2604 Description: libcfs: add cfs_{need,cond}_resched() interface.
2605 Details : libcfs: add cfs_{need,cond}_resched() definition and
2606 implementations for Linux, NT, and liblustre.
2608 Severity : enhancement
2610 Description: More exported tunables for mballoc
2611 Details : Add support for tunable preallocation window and new tunables for
2612 large/small requests
2616 Description: Detect corruption of block bitmap and checking for preallocations
2617 Details : Checks validity of on-disk block bitmap. Also it does better
2618 checking of number of applied preallocations. When corruption is
2619 found, it turns filesystem readonly to prevent further corruptions.
2623 Description: (rw.c:1323:ll_read_ahead_pages()) ASSERTION(page_idx > ria->ria_stoff) failed
2624 Details : Once the unmatched stride IO mode is detected, shrink the stride-ahead
2625 window to 0. If it does hit cache miss, and read-pattern is still
2626 stride-io mode, does not reset the stride window, but also does not
2627 increase the stride window length in this case.
2631 Frequency : only for big-endian servers
2632 Description: Check if system is big-endian while mounting fs with extents feature
2633 Details : Mounting a filesystem with extents feature will fail on big-endian
2634 systems since ext3-based ldiskfs is not supported on big-endian
2635 systems. This can be over-riden with "bigendian_extents" mount option.
2637 Severity : enhancement
2639 Description: The root squash functionality
2640 Details : A security feature, which is to prevent users from being able
2641 to mount lustre on their desktop, run as root, and delete
2642 all of the files in the filesystem. The goal is accomplished by
2643 remapping user id (UID) and group id (GID) of the root user to
2644 a UID and GID specified by the system administartor via Lustre
2645 configuration management server (MGS). The functionality also
2646 allows to specify sets of clients for which the remapping does
2651 Description: Excessive recovery window
2652 Details : With AT enabled, the recovery window can be excessively long (6000+
2653 seconds). To address this problem, we no longer use
2654 OBD_RECOVERY_FACTOR when extending the recovery window (the connect
2655 timeout no longer depends on the service time, it is set to
2656 INITIAL_CONNECT_TIMEOUT now) and clients report the old service
2657 time via pb_service_time.
2661 Description: Watchdog triggered on MDS failover
2662 Details : enable OBD_CONNECT_MDT flag when connecting from the MDS so that
2663 the OSTs know that the MDS "UUID" can be reused for the same export
2664 from a different NID, so we do not need to wait for the export to be
2668 Frequency : rare, only if using MMP with Linux RAID
2670 Description: MMP doesn't work with Linux RAID
2671 Details : While using HA for Lustre servers with Linux RAID, it is possible
2672 that MMP will not detect multiple mounts. To make this work we
2673 need to unplug the device queue in RAID when the MMP block is being
2674 written. Also while reading the MMP block, we should read it from
2675 disk and not the cached one.
2677 Severity : enhancement
2679 Description: open file using fid
2680 Details : A file can be opened using just its fid, like
2681 <mntpt>/.lustre/fid/SEQ:OID:VER - this is needed for HSM and replication
2684 Frequency : Only in RHEL5 when mounting multiple ext3 filesystems
2687 Description: "kmem_cache_create: duplicate cache jbd_4k" error message
2688 Details : add proper locking for creation of jbd_4k slab cache
2692 Description: MMP check in ext3_remount() fails without displaying any error
2693 Details : When multiple mount protection fails during remount, proper error
2696 Severity : enhancement
2698 Description: Allow stripe size to be up to 4G-64k
2699 Details : Fix math logic to allow large stripe sizes.
2703 Description: add check for >8TB ldiskfs filesystems
2704 Details : ext3-based ldiskfs does not support greater than 8TB LUNs.
2705 Don't allow >8TB ldiskfs filesystems to be mounted without
2706 force_over_8tb mount option
2711 Descriptoin: OSS complains "too many blocks" and crashes
2712 Details : In filter_commitrw_write():
2713 1. lquota_chkquota();
2714 2. obdo_from_inode();
2715 3. lquota_pending_commit();
2716 "2." has changed uid and gid in this case, then caused
2717 lquota_chkquota() and lquota_pending_commit() to use
2718 different uid and gid. That is the root of the bug.
2722 Descriptoin: getstripe doesn't report stripe count = -1 sanely
2723 Details : Detect and print out the 'stripe count = -1' case properly.
2727 Description: Out or order replies might be lost on replay
2728 Details : In ptlrpc_retain_replayable_request if we cannot find retained
2729 request with tid smaller then one currently being added, add it
2730 to the start, not end of the list.
2732 --------------------------------------------------------------------------------
2734 2007-08-10 Cluster File Systems, Inc. <info@clusterfs.com>
2736 * Support for kernels:
2737 2.6.5-7.283 (SLES 9),
2738 2.6.9-55.EL (RHEL 4),
2739 2.6.16.46-0.14 (SLES 10),
2740 2.6.18.8 vanilla (kernel.org)
2741 * Client support for unpatched kernels:
2742 (see http://wiki.lustre.org/index.php?title=Patchless_Client)
2743 2.6.16 - 2.6.22 vanilla (kernel.org)
2744 * Due to recently discovered recovery problems, we do not recommend
2745 using patchless RHEL 4 clients with this or any earlier release.
2746 * Recommended e2fsprogs version: 1.39.cfs8
2747 * Note that reiserfs quotas are disabled on SLES 10 in this kernel.
2748 * Starting with this release, the ldiskfs backing filesystem required
2749 by Lustre is now in its own package, lustre-ldiskfs. This package
2750 should be installed. It is versioned separately from Lustre and
2751 may be released separately in future.
2755 Description: block reactivating mgc import until all deactivates complete
2756 Details : Fix race when failing back MDT/MGS to itself (testing)
2759 Frequency : at statup only
2761 Description: mds_lov_synchronize race leads to various problems
2762 Details : simultaneous MDT->OST connections at startup can cause the
2763 sync to abort, leaving the OSC in a bad state.
2765 Severity : enhancement
2767 Description: add optional extra BUILD_VERSION info
2768 Details : add a new environment variable (namely LUSTRE_VERS) which allows
2769 to override the lustre version.
2772 Frequency : 2.6.18 servers only
2774 Description: ll_kern_mount() doesn't release the module reference
2775 Details : The ldiskfs module reference count never drops down to 0
2776 because ll_kern_mount() doesn't release the module reference.
2781 Description: server LBUG when using old ost_num_threads parameter
2782 Details : Accept the old ost_num_threads parameter but warn that it
2783 is deprecated, and fix an off-by-one error that caused an LBUG.
2788 Description: Transient SCSI error results in persistent IO issue
2789 Details : iobuf->dr_error is not reinitialized to 0 between two
2793 Frequency : sometimes when underlying device returns I/O errors
2795 Description: OSTs not going read-only during write failures
2796 Details : OSTs are not remounted read-only when the journal commit threads
2797 get I/O errors because fsfilt_ext3 calls journal_start/stop()
2798 instead of the ext3 wrappers.
2802 Description: poor connect scaling with increasing client count
2803 Details : Don't run filter_grant_sanity_check for more than 100 exports
2804 to improve scaling for large numbers of clients.
2807 Frequency : SLES10 only
2809 Description: sanity-quota.sh quotacheck failed: rc = -22
2810 Details : Quotas cannot be enabled on SLES10.
2813 Frequency : liblustre clients only
2815 Description: getdirentries does not give error when run on compute nodes
2816 Details : getdirentries does not fail when the size specified as an argument
2817 is too small to contain at least one entry
2819 Severity : enhancement
2821 Description: Add LNET router traceability for debug purposes
2822 Details : If a checksum failure occurs with a router as part of the
2823 IO path, the NID of the last router that forwarded the bulk data
2824 is printed so it can be identified.
2829 Description: OST "spontaneously" evicts client; client has imp_pingable == 0
2830 Details : Due to a race condition, liblustre clients were occasionally
2831 evicted incorrectly.
2833 Severity : enhancement
2835 Description: lfs setstripe use optional parameters instead of postional
2838 Severity : enhancement
2840 Description: Nanosecond timestamp support for ldiskfs
2841 Details : The on-disk ldiskfs filesystem has added support for nanosecond
2842 resolution timestamps. There is not yet support for this at
2843 the Lustre filesystem level.
2846 Frequency : during server recovery
2848 Description: MDS failing to send precreate requests due to OSCC_FLAG_RECOVERING
2849 Details : request with rq_no_resend flag not awake l_wait_event if they get
2853 Frequency : nfs export on patchless client
2855 Description: connectathon hang when test nfs export over patchless client
2856 Details : Disconnected dentry cannot be found with lookup, so we do not need
2857 to unhash it or make it invalid
2860 Description: fix llapi_lov_get_uuids() to allow many OSTs to be returned
2861 Details: : Change llapi_lov_get_uuids() to read the UUIDs from /proc instead
2862 of using an ioctl. This allows lfsck for > 160 OSTs to succeed.
2867 Description: open req refcounting wrong on reconnect
2868 Details : If reconnect happened between getting open reply from server and
2869 call to mdc_set_replay_data in ll_file_open, we will schedule
2870 replay for unreferenced request that we are about to free.
2871 Subsequent close will crash in variety of ways.
2872 Check that request is still eligible for replay in
2873 mdc_set_replay_data().
2878 Description: disable writes to filesystem when reading health_check file
2879 Details : the default for reading the health_check proc file has changed
2880 to NOT do a journal transaction and write to disk, because this
2881 can cause reads of the /proc file to hang and block HA state
2882 checking on a healthy but otherwise heavily loaded system. It
2883 is possible to return to the previous behaviour during configure
2884 with --enable-health-write.
2886 Severity : enhancement
2888 Description: 64-bit inode version
2889 Details: : Add a on-disk 64-bit inode version for ext3 to track changes made
2890 to the inode. This will be required for version-based recovery.
2895 Description: MDS fails to start if a duplicate client export is detected
2896 Details : in some rare cases it was possible for a client to connect to
2897 an MDS multiple times. Upon recovery the MDS would detect this
2898 and fail during startup. Handle this more gracefully.
2900 Severity : enhancement
2902 Description: Add -o localflock option to simulate old noflock
2904 Details : This will achieve local-only flock/fcntl locks
2910 Description: log_commit_thread vs filter_destroy race leads to crash
2911 Details : Take import reference before releasing llog record semaphore
2916 Description: Wrong request locking in request set processing
2917 Details : ptlrpc_check_set wrongly uses req->rq_lock for proctect add to
2918 imp_delayed_list, in this place should be used imp_lock.
2921 Frequency : when reconnection
2923 Description: Grant Leak when osc reconnect to OST
2924 Details : When osc reconnect ost, OST(filter) should check whether it
2925 should grant more space to client by comparing fed_grant and
2926 cl_avail_grant, and return the granted space to client instead
2927 of "new granted" space, because client will call osc_init_grant
2928 to update the client grant space info.
2931 Frequency : when client reconnect to OST
2933 Description: Grant Leak when osc do resend and replay bulk write
2934 Details : When osc reconnect to OST, OST(filter)should clear grant info of
2935 bulk write request, because the grant info while be sync between
2936 OSC and OST when reconnect, and we should ignore the grant info
2937 these of resend/replay write req.
2942 Description: Grant space more than available left space sometimes.
2943 Details : When then OST is about to be full, if two bulk writing from
2944 different clients came to OST. Accord the avaliable space of the
2945 OST, the first req should be permitted, and the second one
2946 should be denied by ENOSPC. But if the seconde arrived before
2947 the first one is commited. The OST might wrongly permit second
2948 writing, which will cause grant space > available space.
2951 Frequency : when client is evicted
2953 Description: Grant might be wrongly erased when osc is evicted by OST
2954 Details : when the import is evicted by server, it will fork another
2955 thread ptlrpc_invalidate_import_thread to invalidate the
2956 import, where the grant will be set to 0. While the original
2957 thread will update the grant it got when connecting. So if
2958 the former happened latter, the grant will be wrongly errased
2959 because of this race.
2964 Description: Checking Stale with correct fid
2965 Details : ll_revalidate_it should uses de_inode instead of op_data.fid2
2966 to check whether it is stale, because sometimes, we want the
2967 enqueue happened anyway, and op_data.fid2 will not be initialized.
2969 Severity : enhancement
2971 Description: update patchless client
2972 Details : Add support for patchless client with 2.6.20, 2.6.21 and RHEL 5
2975 Frequency : only with 2.4 kernel
2977 Description: random memory corruption
2978 Details : size of struct ll_inode_info is to big for union inode.u and this
2979 can be cause of random memory corruption.
2984 Description: Memory leak in recovery
2985 Details : Lov_mds_md was not free in an error handler in mds_create_object.
2986 It should also check obd_fail before fsfilt_start, otherwise if
2987 fsfilt_start return -EROFS,(failover mds during mds recovery).
2988 then the req will return with repmsg->transno = 0 and rc = EROFS.
2989 and we met hit the assert LASSERT(req->rq_reqmsg->transno ==
2990 req->rq_repmsg->transno) in ptlrpc_replay_interpret. Fcc should
2991 be freed no matter whether fsfilt_commit success or not.
2994 Frequency : only with huge count clients
2996 Description: Prevents from taking the superblock lock in llap_from_page for
2998 Details : using LL_ORIGIN_REMOVEPAGE origin flag instead of LL_ORIGIN_UNKNOW
2999 for llap_from_page call in ll_removepage prevents from taking the
3000 superblock lock for a soon died page.
3005 Description: Not check open intent error before release open handle
3006 Details : in some rare cases, the open intent error is not checked before
3007 release open handle, which may cause
3008 ASSERTION(open_req->rq_transno != 0), because it tries to release
3009 the failed open handle.
3014 Description: Set cat log bitmap only after create log success.
3015 Details : in some rare cases, the cat log bitmap is set too early. and it
3016 should be set only after create log success.
3020 Description: Accessing a block bevice can re-enable I/O when Lustre is
3021 tearing down a device.
3022 Details : dev_clear_rdonly(bdev) must be called in kill_bdev() instead of
3027 Description: service threads may hog cpus when there are a lot of requests
3028 Details : Insert cond_resched to give other threads a chance to use some CPU
3033 Description: the cat log was not initialized in recovery
3034 Details : When mds(mgs) do recovery, the tgt_count might be zero, so the
3035 unlink log on mds will not be initialized until mds post
3036 recovery. And also in mds post recovery, the unlink log will
3037 initialization will be done asynchronausly, so there will be race
3038 between add unlink log and unlink log initialization.
3042 Description: brw_stats were being printed incorrectly
3043 Details : brw_stats were being printed as log2 but all of them were not
3044 recorded as log2. Also remove some code duplication arising from
3045 filter_tally_{read,write}.
3049 Frequency : rare, only in recovery.
3050 Description: ASSERTION(req->rq_type != LI_POISON) failed
3051 Details : imp_lock should be held while iterating over imp_sending_list for
3052 prevent destroy request after get timeout in ptlrpc_queue_wait.
3056 Description: replay-single.sh test 52 fails
3057 Details : A lock's skiplist need to be cleanup when it being unlinked
3058 from its resource list.
3062 Description: Short directio read returns full requested size rather than
3064 Details : Direct I/O operations should return actual amount of bytes
3065 transferred rather than requested size.
3067 Severity : enhancement
3069 Description: metadata RPC reduction (e.g. for rm performance)
3070 Details : decrease the amount of synchronous RPC between clients and servers
3071 by canceling conflicing lock before the operation on the client
3072 and packing thier handles into the main operation RPC to server.
3074 Severity : enhancement
3076 Description: Async OSC create to avoid the blocking unnecessarily.
3077 Details : If an OST has no remaining object, system will block on the
3078 creation when it needs to create a new object on this OST. Now,
3079 ways use pre-created objects when available, instead of blocking on
3080 an empty osc while others are not empty. If we must block, we
3081 block for the shortest possible period of time.
3085 Description: improve handling recoverable errors
3086 Details : if request processig with error which can be recoverable on server
3087 request should be resend, otherwise page released from cache and
3090 Severity : enhancement
3092 Description: refine locking for avoid write wrong info into lov_objid file
3093 Details : fix possible races with add new target and write/update data in
3096 --------------------------------------------------------------------------------
3098 2007-05-03 Cluster File Systems, Inc. <info@clusterfs.com>
3103 Frequency : on some architectures
3105 Description: 1.6 client sometimes fails to mount from a 1.4 MDT
3106 Details : Uninitialized flags sometimes cause configuration commands to
3110 Frequency : patchless clients only
3112 Description: missing __iget() symbol export
3113 Details : The __iget() symbol export is missing. To avoid the need for
3114 this on patchless clients the deathrow inode reaper is turned
3115 off, and we depend on the VM to clean up old inodes. This
3116 dependency was during via the fix for bug 12181.
3118 --------------------------------------------------------------------------------
3120 2007-04-19 Cluster File Systems, Inc. <info@clusterfs.com>
3122 * CONFIGURATION CHANGE. This version of Lustre WILL NOT
3123 INTEROPERATE with older versions automatically. In many cases a
3124 special upgrade step is needed. Please read the
3125 user documentation before upgrading any part of a live system.
3126 * WIRE PROTOCOL CHANGE from previous 1.6 beta versions. This
3127 version will not interoperate with 1.6 betas before beta5 (1.5.95).
3128 * WARNING: Lustre configuration and startup changes are required with
3129 this release. See https://mail.clusterfs.com/wikis/lustre/MountConf
3133 Severity : enhancement
3135 Description: MountConf
3136 Details : Lustre configuration is now managed via mkfs and mount
3137 commands instead of lmc and lconf. New obd types (MGS, MGC)
3138 are added for dynamic configuration management. See
3139 https://mail.clusterfs.com/wikis/lustre/MountConf for
3142 Severity : enhancement
3144 Description: dynamic OST addition
3145 Details : OSTs can now be added to a live filesystem
3147 Severity : enhancement
3149 Description: startup order invariance
3150 Details : MDTs and OSTs can be started in any order. Clients only
3151 require the MDT to complete startup.
3153 Severity : enhancement
3155 Description: parallel, asynchronous orphan cleanup
3156 Details : orphan cleanup is now performed in separate threads for each
3157 OST, allowing parallel non-blocking operation.
3159 Severity : enhancement
3161 Description: optimized stripe assignment
3162 Details : stripe assignments are now made based on ost space available,
3163 ost previous usage, and OSS previous usage, in order to try
3164 to optimize storage space and networking resources.
3166 Severity : enhancement
3168 Description: Permanently set tunables
3169 Details : All writable /proc/fs/lustre tunables can now be permanently
3170 set on a per-server basis, at mkfs time or on a live system.
3172 Severity : enhancement
3174 Description: Lustre message v2
3175 Details : Add lustre message format v2.
3177 Severity : enhancement
3179 Description: client OST exclusion list
3180 Details : Clients can be started with a list of OSTs that should be
3181 declared "inactive" for known non-responsive OSTs.
3185 Description: SPEC SFS validation failure on NFS v2 over lustre.
3186 Details : Changes the blocksize for regular files to be 2x RPC size,
3187 and not depend on stripe size.
3189 Severity : enhancement
3191 Description: Multiple MD RPCs in flight.
3192 Details : Further unserialise some read-only MDS RPCs - learn about intents.
3193 To avoid overly-overloading MDS, introduce a limit on number of
3194 MDS RPCs in flight for a single client and add /proc controls
3195 to adjust this limit.
3197 Severity : enhancement
3199 Description: client read/write statistics
3200 Details : Add client read/write call usage stats for performance
3201 analysis of user processes.
3202 /proc/fs/lustre/llite/*/offset_stats shows non-sequential
3203 file access. extents_stats shows chunk size distribution.
3204 extents_stats_per_process show chunk size distribution per
3207 Severity : enhancement
3209 Description: mds statistics
3210 Details : Add detailed mds operations statistics in
3211 /proc/fs/lustre/mds/*/stats.
3215 Description: Failure of copying files with lustre special EAs.
3216 Details : Client side always return success for setxattr call for lustre
3217 special xattr (currently only "trusted.lov").
3222 Description: Refcount LNET uuids
3223 Details : The global LNET uuid list grew linearly with every startup;
3224 refcount repeated list entries instead of always adding to
3227 Severity : enhancement
3229 Description: Dynamic service threads
3230 Details : Within a small range, start extra service threads
3231 automatically when the request queue builds up.
3234 Frequency : mixed-endian client/server environments
3236 Description: mixed-endian crashes
3237 Details : The new msg_v2 system had some failures in mixed-endian
3240 Severity : enhancement
3242 Description: Easy OST removal
3243 Details : OSTs can be permanently deactivated with e.g. 'lctl
3244 conf_param lustre-OST0001.osc.active=0'
3246 Severity : enhancement
3248 Description: MGS proc entries
3249 Details : Added basic proc entries for the MGS showing what filesystems
3252 Severity : enhancement
3254 Description: provide MGS failover
3255 Details : Added config lock reacquisition after MGS server failover.
3257 Severity : enhancement
3259 Description: add Linux 2.4 support
3260 Details : Added support for RHEL 2.4.21 kernel for 1.6 servers and clients
3264 Description: a large application tries to do I/O to the same resource and dies
3265 in the middle of it.
3266 Details : Check the req->rq_arrival time after the call to
3267 ost_brw_lock_get(), but before we do anything about
3268 processing it & sending the BULK transfer request. This
3269 should help move old stale pending locks off the queue as
3270 quickly as obd_timeout.
3273 Frequency : when an incorrect nid is specified during startup
3275 Description: ptlrpc connect to non-existant node causes kernel crash
3276 Details : LNET can't be re-entered from an event callback, which
3277 happened when we expire a message after the export has been
3278 cleaned up. Instead, hand the zombie cleanup off to another
3281 Severity : enhancement
3283 Description: plain/inodebits lock performance improvement
3284 Details : Grouping plain/inodebits in granted list by their request modes
3285 and bits policy, thus improving the performance of search through
3289 Frequency : only if OST filesystem is corrupted
3291 Description: client incorrectly hits assertion in ptlrpc_replay_req()
3292 Details : for a short time RPCs with bulk IO are in the replay list,
3293 but replay of bulk IOs is unimplemented. If the OST filesystem
3294 is corrupted due to disk cache incoherency and then replay is
3295 started it is possible to trip an assertion. Avoid putting
3296 committed RPCs into the replay list at all to avoid this issue.
3299 Frequency : liblustre (e.g. catamount) on a large cluster with >= 8 OSTs/OSS
3301 Description: System hang on startup
3302 Details : This bug allowed the liblustre (e.g. catamount) client to
3303 return to the app before handling all startup RPCs. This
3304 could leave the node unresponsive to lustre network traffic
3305 and manifested as a server ptllnd timeout.
3307 Severity : enhancement
3309 Description: Add "/proc/sys/lustre/debug_peer_on_timeout"
3310 Details : liblustre envirable: LIBLUSTRE_DEBUG_PEER_ON_TIMEOUT
3311 boolean to control whether to print peer debug info when a
3312 client's RPC times out.
3315 Frequency : only for kernels with patches from Lustre below 1.4.3
3317 Description: Remove old rdonly API
3318 Details : Remove old rdonly API which unused from at least lustre 1.4.3
3321 Frequency : only for devices with external journals
3323 Description: Set external device read-only also
3324 Details : During a commanded failover stop, we set the disk device
3325 read-only while the server shuts down. We now also set any
3326 external journal device read-only at the same time.
3329 Frequency : when upgrading from 1.4 while trying to change parameters
3331 Description: The wrong (new) MDC name was used when setting parameters for
3332 upgraded MDT's. Also allows changing of OSC (and MDC)
3333 parameters if --writeconf is specified at tunefs upgrade time.
3336 Frequency : when setting specific ost indicies
3338 Description: QOS code breaks on skipped indicies
3339 Details : Add checks for missing OST indicies in the QOS code, so OSTs
3340 created with --index need not be sequential.
3342 Severity : enhancement
3344 Description: Add uninit_groups feature to ldiskfs2 to speed up e2fsck
3345 Details : The uninit_groups feature works in conjunction with the kernel
3346 filesystem code (ldiskfs2 only) and e2fsprogs-1.39-cfs6 to speed
3347 up the pass1 processing of e2fsck. This is a read-only feature
3348 in ldiskfs2 only, so older kernels and current ldiskfs cannot
3349 mount filesystems that have had this feature enabled.
3351 Severity : enhancement
3353 Description: Improve multi-block allocation algorithm to avoid fragmentation
3354 Details : The mballoc3 code (ldiskfs2 only) adds new mechanisms to improve
3355 allocation locality and avoid filesystem fragmentation.
3357 ------------------------------------------------------------------------------
3359 2007-04-01 Cluster File Systems, Inc. <info@clusterfs.com>
3361 * Support for kernels:
3362 2.4.21-47.0.1.EL (RHEL 3)
3363 2.6.5-7.283 (SLES 9)
3364 2.6.9-42.0.10.EL (RHEL 4)
3365 2.6.12.6 vanilla (kernel.org)
3366 2.6.16.27-0.9 (SLES 10)
3367 * Recommended e2fsprogs version: 1.39.cfs5
3369 * Note that reiserfs quotas are disabled on SLES 10 in this kernel
3373 Frequency : occasional, depends on client load and configuration
3374 Bugzilla : 12181, 12203
3375 Description: data loss for recently-modified files
3377 Details : In some cases it is possible that recently written or created
3378 files may not be written to disk in a timely manner (this should
3379 normally be within 30s unless client IO load is very high).
3380 The problem appears as zero-length files or files that are a
3381 multiple of 1MB in size after a client crash or client eviction
3382 that are missing data at the end of the file.
3384 This problem is more likely to be hit on clients where files are
3385 repeatedly created and unlinked in the same directory, clients
3386 have a large amount of RAM, have many CPUs, the filesystem has
3387 many OSTs, the clients are rebooted frequently, and/or the files
3388 are not accessed by other nodes after being written.
3390 The presence of the problem can be detected by looking at
3391 /proc/sys/fs/inode-state. If the first number (nr_inodes) is
3392 smaller than the second (nr_unused) then dirty files will not
3393 be flushed automatically to disk. "sync; sleep 10" should be
3394 run several times on the node before unmounting it to update
3395 Lustre (this is also safe to run on nodes without this problem).
3397 There is also a related kernel bug in the RHEL4 4 2.6.9 kernel
3398 that can cause this same problem, so customers using that kernel
3399 also need to update the kernel in addition to Lustre. In order
3400 to properly fix this bug, the RHEL3 2.4.21 kernel is also updated.
3402 It is normal that files written just before a client crash (less
3403 than 30s) may not yet have been flushed to disk, even for local
3407 Frequency : frequent on thin XT3 nodes
3409 Description: UUID collision on thin XT3 Linux nodes
3410 Details : UUIDs on Compute Node Linux XT3 nodes were not generated
3411 randomly, since we relied on an insufficiently-seeded PRNG.
3416 Description: OSS hangs after "All ost request buffers busy"
3417 Details : A deadlock between quota and journal operations caused OSS
3418 hangs after printing "All ost request buffers busy."
3421 Frequency : always on liblustre builds
3423 Description: Cleanup compiler warnings on liblustre
3426 Frequency : always on liblustre builds on XT3
3428 Description: LC_CONFIG_CDEBUG don't run while build liblustre on XT3.
3432 Description: Addition of EXT3_FEATURE_RO_COMPAT_DIR_NLINKS flag for
3433 > 32000 subdirectories
3434 Details : Add EXT3_FEATURE_RO_COMPAT_DIR_NLINK flag to
3435 EXT3_FEATURE_RO_COMPAT_SUPP. This flag will be set whenever
3436 subdirectory count crosses 32000. This will aid e2fsck to
3437 correctly handle more than 32000 subdirectories.
3440 Frequency : liblustre (e.g. catamount) on a large cluster with >= 8 OSTs/OSS
3442 Description: System hang on startup
3443 Details : This bug allowed the liblustre (e.g. catamount) client to
3444 return to the app before handling all startup RPCs. This
3445 could leave the node unresponsive to lustre network traffic
3446 and manifested as a server ptllnd timeout.
3448 Severity : enhancement
3450 Description: Add "/proc/sys/lustre/debug_peer_on_timeout"
3451 (liblustre envirable: LIBLUSTRE_DEBUG_PEER_ON_TIMEOUT)
3452 boolean to control whether to print peer debug info when a
3453 client's RPC times out.
3458 Description: make O_SYNC working on 2.6 kernels
3459 Details : 2.6 kernels use different method for mark pages for write,
3460 so need add a code to lustre for O_SYNC work.
3465 Description: Failure to close file and release space on NFS
3466 Details : Put inode details into lock acquired in ll_intent_file_open.
3467 Use mdc_intent_lock in ll_intent_open to properly
3468 detect all kind of errors unhandled by mdc_enqueue.
3473 Description: proc file read during shutdown sometimes raced obd removal,
3475 Details : Add lock to prevent obd access after proc file removal.
3478 Frequency : Only for files larger than 4GB on 32-bit clients.
3480 Description: improperly doing page alignment of locks
3481 Details : Modify lustre core code to use CFS_PAGE_* defines instead of
3482 PAGE_*. Make CFS_PAGE_MASK a 64-bit mask.
3487 Description: RPCs being resent when they shouldn't be
3488 Details : Some RPCs that should not be resent are being resent. This
3489 can cause inconsistencies in the RPC state machine. Do not
3490 resend such requests.
3493 Frequency : rare, only with NFS export
3495 Description: Crash on NFS re-export node
3496 Details : under very unusual load conditions an assertion is hit in
3497 ll_intent_file_open()
3500 Frequency : only if OST filesystem is corrupted
3502 Description: client incorrectly hits assertion in ptlrpc_replay_req()
3503 Details : for a short time RPCs with bulk IO are in the replay list,
3504 but replay of bulk IOs is unimplemented. If the OST filesystem
3505 is corrupted due to disk cache incoherency and then replay is
3506 started it is possible to trip an assertion. Avoid putting
3507 committed RPCs into the replay list at all to avoid this issue.
3512 Description: large O_DIRECT requests fail under memory pressure/fragmentation
3513 Details : Large single O_DIRECT read and write calls can fail to allocate
3514 a sufficiently large buffer to process the request. In case of
3515 allocation failure the allocation is retried with a smaller
3516 buffer and broken into smaller requests.
3518 Severity : enhancement
3520 Description: Add -o localflock option to simulate old noflock behaviour.
3521 Details : This will achieve local-only flock/fcntl locks coherentness.
3526 Description: versioning check is incomplete
3527 Details : Checking the version difference of client vs. server, report
3528 error if the gap is too big.
3533 Description: add support PG_writeback bit
3534 Details : add support for PG_writeback bit for Lustre, for more carefull
3535 work with page cache in 2.6 kernel. This also fix some deadlocks
3536 and remove hack for work O_SYNC with 2.6 kernel.
3538 Severity : enhancement
3540 Description: Add uninit_groups feature to ldiskfs2 to speed up e2fsck
3541 Details : The uninit_groups feature works in conjunction with the kernel
3542 filesystem code (ldiskfs2 only) and e2fsprogs-1.39-cfs6 to speed
3543 up the pass1 processing of e2fsck. This is a read-only feature
3544 in ldiskfs2 only, so older kernels and current ldiskfs cannot
3545 mount filesystems that have had this feature enabled.
3547 Severity : enhancement
3549 Description: Improve multi-block allocation algorithm to avoid fragmentation
3550 Details : The mballoc3 code (ldiskfs2 only) adds new mechanisms to improve
3551 allocation locality and avoid filesystem fragmentation.
3553 ------------------------------------------------------------------------------
3555 2007-02-09 Cluster File Systems, Inc. <info@clusterfs.com>
3557 * Support for kernels:
3558 2.6.9-42.0.3.EL (RHEL 4)
3559 2.6.5-7.276 (SLES 9)
3560 2.4.21-47.0.1.EL (RHEL 3)
3561 2.6.12.6 vanilla (kernel.org)
3562 2.6.16.21-0.8 (SLES10)
3563 * Recommended e2fsprogs version: 1.39.cfs2-0
3565 * The backwards-compatible /proc/sys/portals symlink has been removed
3566 in this release. Before upgrading, please ensure that you change
3567 any configuration scripts or /etc/sysctl.conf files that access
3568 /proc/sys/portals/* or sysctl portals.* to use the corresponding
3569 entry in /proc/sys/lnet or sysctl lnet.*. This change can be made
3570 in advance of the upgrade on any system running Lustre 1.4.6 or
3571 newer, since /proc/sys/lnet was added in that version.
3572 * Note that reiserfs quotas are disabled on SLES 10 in this kernel
3576 Frequency : only when quota is used
3578 Description: avoid scanning export list for quota master
3579 Details : Change the algorithms to avoid scanning export list in order
3580 to improve the efficiency.
3583 Frequency : MDS failover only, very rarely
3585 Description: "went back in time" messages on mds failover
3586 Details : The greatest transno may be lost when the current operation
3587 finishes with an error (transno==0) and the client's last_rcvd
3588 record is over-written. Save the greatest transno in the
3589 mds_last_transno for this case.
3592 Frequency : always for specific kernels and striping counts
3594 Description: client may get "Matching packet too big" without ACL support
3595 Details : Clients compiled without CONFIG_FS_POSIX_ACL get an error message
3596 when trying to access files in certain configurations. The
3597 clients should in fact be denied when mounting because they do
3598 not understand ACLs.
3601 Frequency : Cray XT3 with more than 4000 clients and multiple jobs
3603 Description: many clients connecting with IO in progress causes connect timeouts
3604 Details : Avoid synchronous journal commits to avoid delays caused by many
3605 clients connecting/disconnecting when bulk IO is in progress.
3606 Queue liblustre connect requests on OST_REQUEST_PORTAL instead of
3607 OST_IO_PORTAL to avoid delays behind potentially many pending
3611 Frequency : occasionally with multiple writers to a single file
3613 Description: shared writes to file may result in wrong size reported by stat()
3614 Details : Allow growing of kms when extent lock is cancelled
3617 Frequency : always with random mmap IO to multi-striped file
3619 Description: mmap write might be lost if we are writing to a 'hole' in stripe
3620 Details : Only if the hole is at the end of OST object so that kms is too
3621 small. Fix is to increase kms accordingly in ll_nopage.
3624 Frequency : rare, only if OST filesystem is inconsistent with MDS filesystem
3626 Description: writes to a missing object would leak memory on the OST
3627 Details : If there is an inconsistency between the MDS and OST filesystems,
3628 such that the MDS references an object that doesn't exist, writes
3629 to that object will leak memory due to incorrect cleanup in the
3630 error handling path, eventually running out of memory on the OST.
3635 Description: Creating too long symlink causes lustre errors
3636 Details : Check symlink and name lengths before sending requests to MDS.
3639 Frequency : only if flock is enabled (not on by default)
3641 Description: posix locks not released on fd closure on 2.6.9+
3642 Details : We failed to add posix locks to list of inode locks on 2.6.9+
3643 kernels, this caused such locks not to be released on fd close and
3644 then assertions on fs unmount about still used locks.
3647 Frequency : MDS failover only, very rarely
3649 Description: clients may get ASSERTION(granted_lock != NULL)
3650 Details : When request was taking a long time, and a client was resending
3651 a getattr by name lock request. The were multiple lock requests
3652 with the same client lock handle and
3653 mds_getattr_name->fixup_handle_for_resent_request found one of the
3654 lock handles but later failed with ASSERTION(granted_lock != NULL).
3659 Description: handle->h_buffer_credits > 0, assertion failure
3660 Details : h_buffer_credits is zero after truncate, causing assertion
3661 failure. This patch extends the transaction or creates a new
3665 Frequency : NFS re-export or patchless client
3666 Bugzilla : 11179, 10796
3667 Description: Crash on NFS re-export node (__d_move)
3668 Details : We do not want to hash the dentry if we don't have a lock.
3669 But if this dentry is later used in d_move, we'd hit uninitialised
3670 list head d_hash, so we just do this to init d_hash field but
3671 leave dentry unhashed.
3674 Frequency : NFS re-export or patchless client
3676 Description: NFS exports has problem with symbolic link
3677 Details : lustre client didn't properly install dentry when re-exported
3678 to NFS or running patchless client.
3681 Frequency : NFS re-export or patchless client
3683 Description: Various nfs/patchless fixes.
3684 Details : fixes reuse disconected alias for lookup process - this fixes
3685 warning "find_exported_dentry: npd != pd",
3686 fix permission error with open files at nfs.
3687 fix apply umask when do revalidate.
3690 Frequency : occasional
3692 Description: Crash on NFS re-export node
3693 Details : calling clear_page() on the wrong pointer triggered oops in
3694 generic_mapping_read().
3697 Frequency : rarely, using O_DIRECT IO
3699 Description: unaligned directio crashes client with LASSERT
3700 Details : check for unaligned buffers before trying any requests.
3703 Frequency : rarely, using CFS RAID5 patches in non-standard kernel series
3705 Description: stale data returned from RAID cache
3706 Details : If only a small amount of IO is done to the RAID device before
3707 reading it again it is possible to get stale data from the RAID
3708 cache instead of reading it from disk.
3711 Frequency : always for sles10 kernel
3713 Description: sles10 support
3714 Details : ll_follow_link: compile fixes and using of nd_set_link
3715 under newer kernels.
3718 Frequency : depends on arch, kernel and compiler version, always on sles10
3721 Description: recursive or deep enough symlinks cause stack overflow
3722 Details : getting rid of large stack-allocated variable in
3726 Frequency : depends on hardware
3728 Description: lustre write performance loss in the SLES10 kernel
3729 Details : the performance loss is caused by using of write barriers in the
3730 ext3 code. The SLES10 kernel turns barrier support on by
3731 default. The fix is to undo that change for ldiskfs.
3733 ------------------------------------------------------------------------------
3735 2006-12-09 Cluster File Systems, Inc. <info@clusterfs.com>
3737 * Support for kernels:
3738 2.6.9-42.0.3EL (RHEL 4)
3739 2.6.5-7.276 (SLES 9)
3740 2.4.21-47.0.1.EL (RHEL 3)
3741 2.6.12.6 vanilla (kernel.org)
3745 Frequency : quota enabled and large files being deleted
3747 Description: releasing more than 4GB of quota at once hangs OST
3748 Details : If a user deletes more than 4GB of files on a single OST it
3749 will cause the OST to spin in an infinite loop. Release
3750 quota in < 4GB chunks, or use a 64-bit value for 1.4.7.1+.
3755 Description: statfs data retrieved from /proc may be stale or zero
3756 Details : When reading per-device statfs data from /proc, in the
3757 {kbytes,files}_{total,free,avail} files, it may appear
3758 as zero or be out of date.
3761 Frequency : systems with MD RAID1 external journal devices
3763 Description: lconf's call to blkid is confused by RAID1 journal devices
3764 Details : Use the "blkid -l" flag to locate the MD RAID device instead
3765 of returning all block devices that match the journal UUID.
3768 Frequency : always, for aggregate stripe size over 4GB
3770 Description: "lfs setstripe" fails assertion when setting 4GB+ stripe width
3771 Details : Using "lfs setstripe" to set stripe size * stripe count over 4GB
3772 will fail the kernel with "ASSERTION(lsm->lsm_xfersize != 0)"
3775 Frequency : always if "lfs find" used on a local file/directory
3777 Description: "lfs find" segfaults if used on a local file/directory
3778 Details : The case where a directory component was not specified wasn't
3779 handled correctly. Handle this properly.
3782 Frequency : always on ppc64
3784 Description: the write to an ext3 filesystem mounted with mballoc got stuck
3785 Details : ext3_mb_generate_buddy() uses find_next_bit() which does not
3786 perform endianness conversion.
3789 Frequency : rarely (truncate to non-zero file size after write under load)
3790 Bugzilla : 10730, 10687
3791 Description: Files padded with zeros to next 4K multiple
3792 Details : With filesystems mounted using the "extents" option (2.6 kernels)
3793 it is possible that files that are truncated to a non-zero size
3794 immediately after being written are filled with zero bytes beyond
3795 the truncated size. No file data is lost.
3797 Severity : enhancement
3799 Description: Allow recovery/failover for liblustre clients.
3800 Details : liblustre clients were unaware of failover configurations until
3803 Severity : enhancement
3805 Description: user file locks should fail when not mounting with flock option
3806 Details : Set up an error-returning stub in ll_file_operations.lock field
3807 to prevent incorrect behaviour when client is mounted without
3808 flock option. Also, set up properly f_op->flock field for
3812 Frequency : always on ia64
3814 Description: "lfs df" loops on printing out MDS statfs information
3815 Details : The obd_ioctl_data was not initialized and in some systems
3816 this caused a failure during the ioctl that did not return
3817 an error. Initialize the struct and return an error on failure.
3820 Frequency : SLES 9 only
3822 Description: Error of copying files with lustre special EAs as root
3823 Details : Client side always return success for setxattr call for lustre
3824 special xattr (currently only "trusted.lov").
3827 Frequency : rarely on clusters with both ia64+i386 clients
3829 Description: ia64+i686 clients doing shared IO on the same file may LBUG
3830 Details : In rare cases when both ia64+i686 (or other mixed-PAGE_SIZE)
3831 clients are doing concurrent writes to the same file it is
3832 possible that the ia64 clients may LASSERT because the OST
3833 extent locks are not PAGE_SIZE aligned. Ensure that grown
3834 locks are always aligned on the request boundary.
3837 Frequency : specific use, occasional
3839 Description: Overwriting in use executable truncates on-disk binary image
3840 Details : If one node attempts to overwrite an executable in use by
3841 another node, we now correctly return ETXTBSY instead of
3842 truncating the file.
3844 Severity : enhancement
3846 Description: Async OSC create to avoid the blocking unnecessarily.
3847 Details : If an OST has no remaining object, system will block on the creating
3848 when need to create a new object on this OST. Now, ways use
3849 pre-created objects when available, instead of blocking on an
3850 empty osc while others are not empty. If we must block, we block
3851 for the shortest possible period of time.
3856 Description: chmod on Lustre root is propagated to other clients
3857 Details : Re-validate root's dentry in ll_lookup_it to avoid having it
3858 invalid by the follow_mount time.
3861 Frequency : liblustre clients only
3863 Description: Race in 'instant cancel' lock handling could lead to such locks
3864 never to be granted in case of SMP MDS
3865 Details : Do not destroy not yet granted but cbpending locks in
3869 Frequency : replay/resend of open
3871 Description: non null lock assetion failure in mds_intent_policy
3872 Details : Trying to replay/resend lockless open requests resulted in
3873 mds_open() returning 0 with no lock. Now it sets a flag if
3874 a lock is going to be returned.
3876 Severity : enhancement
3878 Description: Checksum enhancements
3879 Details : New checksum enhancements allow for resending RPCs that failed
3882 Severity : enhancement
3884 Description: Tunables on number of dirty pages in cacche
3885 Details : Allow to set limit on number of dirty pages cached.
3890 Description: client crash on unmount - lock still has references
3891 Details : In some error handling cases it was possible to leak a lock
3892 reference on a client while accessing a file. This was not
3893 harmful to the client during operation, but would cause the
3894 client to crash when the filesystem is unmounted.
3897 Frequency : specific case, rare
3899 Description: ETXTBSY on mds though file not in use by client
3900 Details : ETXTBSY is no longer incorrectly returned when attempting to
3901 chmod or chown a directory that the user previously tried to
3902 execute or a currently-executing binary.
3905 Frequency : extremely rare except on liblustre-based clients
3907 Description: Lustre space not freed when files are deleted
3908 Details : Clean up open-unlinked files after client eviction. Previously
3909 the unlink was skipped and the files remained as orphans.
3914 Description: OST failure "would be an LBUG" in waiting_locks_callback()
3915 Details : In some cases it was possible to send a blocking callback to
3916 a client doing a glimpse, even though that client didn't get
3917 a lock granted. When the glimpse lock is cancelled on the OST
3918 the freed lock is left on the waiting list and corrupted the list.
3921 Frequency : all core dumps
3923 Description: Broke core dumps to lustre
3924 Details : Negative dentry may be unhashed if parent does not have UPDATE
3925 lock, but some callers, e.g. do_coredump, expect dentry to be
3926 hashed after successful create, hash it in ll_create_it.
3928 ------------------------------------------------------------------------------
3930 2006-09-13 Cluster File Systems, Inc. <info@clusterfs.com>
3932 * Support for kernels:
3933 2.6.9-42.0.2.EL (RHEL 4)
3934 2.6.5-7.276 (SLES 9)
3935 2.4.21-40.EL (RHEL 3)
3936 2.6.12.6 vanilla (kernel.org)
3940 Frequency : always on RHEL 3
3942 Description: Number of open files grows over time
3943 Details : The number of open files grows over time, whether or not
3944 Lustre is started. This was due to a filp leak introduced
3945 by one of our kernel patches.
3947 ------------------------------------------------------------------------------
3949 08-20-2006 Cluster File Systems, Inc. <info@clusterfs.com>
3951 * Support for kernels:
3952 2.6.9-42.EL (RHEL 4)
3953 2.6.5-7.276 (SLES 9)
3954 2.4.21-40.EL (RHEL 3)
3955 2.6.12.6 vanilla (kernel.org)
3960 Bugzilla : 5719, 9635, 9792, 9684
3961 Description: OST (or MDS) trips assertions in (re)connection under heavy load
3962 Details : If a server is under heavy load and cannot reply to new
3963 connection requests before the client resends the (re)connect,
3964 the connection handling code can behave badly if two service
3965 threads are concurrently handing separate (re)connections from
3966 the same client. Add better locking to the connection handling
3967 code, and ensure that only a single connection will be processed
3968 for a given client UUID, even if the lock is dropped.
3970 Severity : enhancement
3972 Description: add TCP zero-copy support to kernel
3973 Details : Add support to the kernel TCP stack to allow zero-copy bulk
3974 sends if the hardware supports scatter-gather and checksumming.
3975 This allows socklnd to do client-write and server-read more
3976 efficiently and reduce CPU utilization from skbuf copying.
3979 Frequency : only if NFS exporting from client
3981 Description: NULL pointer deref in ll_iocontrol() if chattr mknod file
3982 Details : If setting attributes on a file created under NFS that had
3983 never been opened it would be possible to oops the client
3984 if the file had no objects.
3987 Frequency : always for liblustre
3989 Description: liblustre client does MDS+OSTs setattr RPC for each write
3990 Details : When doing a write from a liblustre client, the client
3991 incorrectly issued an RPC to the MDS and each OST the file was
3992 striped over in order to update the timestamps. When writing
3993 with small chunks and many clients this could overwhelm the MDS
3994 with RPCs. In all cases it would slow down the write because
3995 these RPCs are unnecessary.
3997 Severity : enhancement
3999 Description: allow number of MDS service threads to be changed at module load
4000 Details : It is now possible to change the number of MDS service threads
4001 running. Adding "options mds mds_num_threads={N}" to the MDS's
4002 /etc/modprobe.conf will set the number of threads for the next
4003 time Lustre is restarted (assuming the "mds" module is also
4004 reloaded at that time). The default number of threads will
4005 stay the same, 32 for most systems.
4010 Description: OST crash if filesystem is unformatted or corrupt
4011 Details : If an OST is started on a device that has never been formatted
4012 or if the filesystem is corrupt and cannot even mount then the
4013 error handling cleanup routines would dereference a NULL pointer.
4018 Description: NULL pointer deref in llap_from_page.
4019 Details : get_cache_page_nowait can return a page with NULL (or otherwise
4020 incorrect) mapping if the page was truncated/reclaimed while it was
4021 searched for. Check for this condition and skip such pages when
4022 doing readahead. Introduce extra check to llap_from_page() to
4023 verify page->mapping->host is non-NULL (so page is not anonymous).
4026 Frequency : Sometimes when using sys_sendfile
4028 Description: "page not covered by a lock" warnings from ll_readpage
4029 Details : sendfile called ll_readpage without right page locks present.
4030 Now we introduced ll_file_sendfile that does necessary locking
4031 around call to generic_file_sendfile() much like we do in
4035 Frequency : with certain MDS communication failures at client mount time
4037 Description: NULL pointer deref after failed client mount
4038 Details : a client connection request may delayed by the network layer
4039 and not be sent until after the PTLRPC layer has timed out the
4040 request. If the client fails the mount immediately it will try
4041 to clean up before the network times out the request. Add a
4042 reference from the request import to the obd device and delay
4043 the cleanup until the network drops the request.
4046 Frequency : occasionally during client (re)connect
4048 Description: assertion failure during client (re)connect
4049 Details : processing a client connection request may be delayed by the
4050 client or server longer than the client connect timeout. This
4051 causes the client to resend the connection request. If the
4052 original connection request is replied in this interval, the
4053 client may trip an assertion failure in ptlrpc_connect_interpret()
4054 which thought it would be the only running connect process.
4057 Frequency : only with obd_echo servers and clients that are rebooted
4059 Description: kernel BUG accessing uninitialized data structure
4060 Details : When running an obd_echo server it did not start the ping_evictor
4061 thread, and when a client was evicted an uninitialized data
4062 structure was accessed. Start the ping_evictor in the RPC
4063 service startup instead of the OBD startup.
4065 Severity : enhancement
4066 Bugzilla : 10193 (patchless)
4067 Description: Remove dependency on various unexported kernel interfaces.
4068 Details : No longer need reparent_to_init, exit_mm, exit_files,
4069 sock_getsockopt, filemap_populate, FMODE_EXEC, put_filp.
4072 Frequency : rare (only users of deprecated and unsupported LDAP config)
4074 Description: write_conf for zeroconf mount queried LDAP incorrectly for client
4075 Details : LDAP apparently contains 'lustreName' attributes instead of
4076 'name'. A simple remapping of the name is sufficient.
4079 Frequency : rare (only with non-default dump_on_timeout debug enabled)
4081 Description: waiting_locks_callback trips kernel BUG if client is evicted
4082 Details : Running with the dump_on_timeout debug flag turned on makes
4083 it possible that the waiting_locks_callback() can try to dump
4084 the Lustre kernel debug logs from an interrupt handler. Defer
4085 this log dumping to the expired_lock_main() thread.
4087 Severity : enhancement
4089 Description: Support NFS exporting on 2.6 kernels.
4090 Details : Implement non-rawops metadata methods for NFS server to use without
4091 changing NFS server code.
4094 Frequency : very rare (synthetic metadata workload only)
4096 Description: two racing renames might cause an MDS thread to deadlock
4097 Details : Running the "racer" program may cause one MDS thread to rename
4098 a file from being the source of a rename to being the target of
4099 a rename at exactly the same time that another thread is doing
4100 so, and the second thread has already enqueued these locks after
4101 doing a lookup of the target and is trying to relock them in
4102 order. Ensure that we don't try to re-lock the same resource.
4105 Frequency : only very large systems with liblustre clients
4107 Description: slow eviction of liblustre clients with the "evict_by_nid" RPC
4108 Details : Use asynchronous set_info RPCs to send the "evict_by_nid" to
4109 all OSTs in parallel. This allows the eviction of stale liblustre
4110 clients to proceed much faster than if they were done in series,
4111 and also offers similar improvements for other set_info RPCs.
4116 Description: excessive CPU usage during initial read phase on client
4117 Details : During the initial read phase on a client, it would agressively
4118 retry readahead on the file, consuming too much CPU and impacting
4119 performance (since 1.4.5.8). Improve the readahead algorithm
4120 to avoid this, and also improve some other common cases (read
4121 of small files in particular, where "small" is files smaller than
4122 /proc/fs/lustre/llite/*/max_read_ahead_whole_mb, 2MB by default).
4127 Description: MDS crash when receiving packet with unknown intent.
4128 Details : Do not LBUG in unknown intent case, just return -EFAULT
4130 Severity : enhancement
4131 Bugzilla : 9293, 9385
4132 Description: MDS RPCs are serialised on client. This is unnecessary for some.
4133 Details : Do not serialize getattr (non-intent version) and statfs.
4136 Frequency : occasional, when OST network is overloaded/intermittent
4138 Description: client evicted by OST after bulk IO timeout
4139 Details : If a client sends a bulk IO request (read or write) the OST
4140 may evict the client if it is unresposive to its data GET/PUT
4141 request. This is incorrect if the network is overloaded (takes
4142 too long to transfer the RPC data) or dropped the OST GET/PUT
4143 request. There is no need to evict the client at all, since
4144 the pinger and/or lock callbacks will handle this, and the
4145 client can restart the bulk request.
4148 Frequency : Always when mmapping file with no objects
4150 Description: client crashes when mmapping file with no objects
4151 Details : Check that we actually have objects in a file before doing any
4152 operations on objects in ll_vm_open, ll_vm_close and
4158 Description: Request leak when working with deleted CWD
4159 Details : Introduce advanced request refcount tracking for requests
4160 referenced from lustre intent.
4162 Severity : Enhancement
4164 Description: Cache open file handles on client.
4165 Details : MDS now will return special lock along with openhandle, if
4166 requested and client is allowed to hold openhandle, even if unused,
4167 until such a lock is revoked. Helps NFS a lot, since NFS is opening
4168 closing files for every read/write openration.
4170 Severity : Enhancement
4172 Description: Cache open negative dentries on client when possible.
4173 Details : Guard negative dentries with UPDATE lock on parent dir, drop
4174 negative dentries on lock revocation.
4179 Description: Remounting a client read-only wasn't possible with a zconf mount
4180 Details : It wasn't possible to remount a client read-only with llmount.
4182 Severity : enhancement
4183 Description: Include MPICH 1.2.6 Lustre ADIO interface patch
4184 Details : In lustre/contrib/ or /usr/share/lustre in RPM a patch for
4185 MPICH is included to add Lustre-specific ADIO interfaces.
4186 This is based closely on the UFS ADIO layer and only differs
4187 in file creation, in order to allow the OST striping to be set.
4188 This is user-contributed code and not supported by CFS.
4193 Description: extended inode attributes (immutable, append-only) work improperly
4194 when 2.4 and 2.6 kernels are used on client/server or vice versa
4195 Details : Introduce kernel-independent values for these flags.
4197 Severity : enhancement
4200 Description: Allow fractional MB tunings for lustre in /proc/ filesystem.
4201 Details : Many of the /proc/ tunables can only be tuned at a megabyte
4202 granularity. Now, Fractional MB granularity is be supported,
4203 this is very useful for low memory system.
4205 Severity : enhancement
4207 Description: Getattr by fid
4208 Details : Getting a file attributes by its fid, obtaining UPDATE|LOOKUP
4209 locks, avoids extra getattr rpc requests to MDS, allows '/' to
4210 have locks and avoids getattr rpc requests for it on every stat.
4213 Frequency : Always, for filesystems larger than 2TB
4215 Description: ldiskfs crash at mount for filesystem larger than 2TB with mballoc
4216 Details : Kenrel kmalloc limits allocations to 128kB and this prevents
4217 filesystems larger than 2TB to be mounted with mballoc enabled.
4220 Frequency : Always, for 32-bit kernel without CONFIG_LBD and filesystem > 2TB
4222 Description: ldiskfs crash at mount for filesystem larger than 2TB with mballoc
4223 Details : If a 32-bit kernel is compiled without CONFIG_LBD enabled and a
4224 filesystems larger than 2TB is mounted then the kernel will
4225 silently corrupt the start of the filesystem. CONFIG_LBD is
4226 enabled for all CFS-supported kernels, but the possibility of
4227 this happening with a modified kernel config exists.
4229 Severity : enhancement
4231 Description: add client O_DIRECT support for 2.6 kernels
4232 Details : It is now possible to do O_DIRECT reads and writes to files
4233 in the Lustre client mountpoint on 2.6 kernel clients.
4235 Severity : enhancement
4237 Description: parallel glimpse, setattr, statfs, punch, destroy requests
4238 Details : Sends glimpse, setattr, statfs, punch, destroy requests to OSTs in
4239 parallel, not waiting for response from every OST before sending
4240 a rpc to the next OST.
4245 Description: setattr vs write race when updating file timestamps
4246 Details : Client processes that update a file timestamp into the past
4247 right after writing to the file (e.g. tar) it is possible that
4248 the updated file modification time can be reset to the current
4249 time due to a race between processing the setattr and write RPC.
4251 Severity : enhancement
4253 Description: Bring 'lfs find' closer in line with regular Linux find.
4254 Details : lfs find util supports -atime, -mtime, -ctime, -maxdepth, -print,
4255 -print0 options and obtains all the needed info through the lustre
4258 Severity : enhancement
4260 Description: support up to 1024 configured devices on one node
4261 Details : change obd_dev array from statically allocated to dynamically
4262 allocated structs as they are first used to reduce memory usage
4267 Description: Flush dirty partially truncated pages during truncate
4268 Details : Immediatelly flush partially truncated pages in filter_setattr,
4269 this way we completely avoid having any pages in page cache on OST
4270 and can retire ugly workarounds during writes to flush such pages.
4275 Description: i_sem vs transaction deadlock in mds_obd_destroy during unlink.
4276 Details : protect inode from truncation within vfs_unlink() context
4277 just take a reference before calling vfs_unlink() and release it
4278 when parent's i_sem is free.
4283 Description: last_id value checked outside lock on OST caused LASSERT failure
4284 Details : If there were multiple MDS->OST object precreate requests in
4285 flight, it was possible that the OST's last object id was checked
4286 outside a lock and incorrectly tripped an assertion. Move checks
4287 inside locks, and discard old precreate requests.
4290 Frequency : always, if extents are used on OSTs
4292 Description: index ei_leaf_hi (48-bit extension) is not zeroed in extent index
4293 Details : OSTs using the extents format would not zero the high 16 bits of
4294 the index physical block number. This is not a problem for any
4295 OST filesystems smaller than 16TB, and no kernels support ext3
4296 filesystems larger than 16TB yet. This is fixed in 1.4.7 (all
4297 new/modified files) and can be fixed for existing filesystems
4298 with e2fsprogs-1.39-cfs1.
4303 Description: import connection selection may be incorrect if timer wraps
4304 Details : Using a 32-bit jiffies timer with HZ=1000 may cause backup
4305 import connections to be ignored if the 32-bit jiffies counter
4306 wraps. Use a 64-bit jiffies counter.
4309 Frequency : very large clusters immediately after boot
4311 Description: LNET request buffers exhausted under heavy short-term load
4312 Details : If a large number of client requests are generated on a service
4313 that has previously never seen so many requests it is possible
4314 that the request buffer growth cannot keep up with the spike in
4315 demand. Instead of dropping incoming requests, they are held in
4316 the LND until the RPC service can accept more requests.
4319 Frequency : Sometimes during replay
4321 Description: Assertion failure in ll_local_open after replay.
4322 Details : If replay happened on an open request reply before we were able
4323 to set replay handler, reply will become not swabbed tripping the
4324 assertion in ll_local_open. Now we set the handler right after
4325 recognising of open request
4328 Frequency : very rare
4330 Description: kernel reports "badness in vsnprintf"
4331 Details : Reading from the "recovery_status" /proc file in small chunks
4332 may cause a negative length in lprocfs_obd_rd_recovery_status()
4333 call to vsnprintf() (which is otherwise harmless). Exit early
4334 if there is no more space in the output buffer.
4336 Severity : enhancement
4338 Description: clear OBD RPC statistics by writing to them
4339 Details : It is now possible to clear the OBD RPC statistics by writing
4340 to the "stats" file.
4345 Description: Client mtime is not the same on different clients after utimes
4346 Details : In some cases, the client was using the utimes() syscall on
4347 a file cached on another node. The clients now validate the
4348 ctime from the MDS + OSTs to determine which one is right.
4353 Description: Inability to activate failout mode
4354 Details : lconf script incorrectly assumed that in pythong string's numeric
4355 value is used in comparisons.
4358 Frequency : always with multiple stripes per file
4360 Description: Inefficient object allocation for mutli-stripe files
4361 Details : When selecting which OSTs to stripe files over, for files with
4362 a stripe count that divides evenly into the number of OSTs,
4363 the MDS is always picking the same starting OST for each file.
4364 Return the OST selection heuristic to the original design.
4369 Description: mount failures may take full timeout to return an error
4370 Details : Under some heavy load conditions it is possible that a
4371 failed mount can wait for the full obd_timeout interval,
4372 possibly several minutes, before reporting an error.
4373 Instead return an error as soon as the status is known.
4375 Frequency : quota enabled and large files being deleted
4377 Description: releasing more than 4GB of quota at once hangs OST
4378 Details : If a user deletes more than 4GB of files on a single OST it
4379 will cause the OST to spin in an infinite loop. Release
4380 quota in < 4GB chunks, or use a 64-bit value for 1.4.7.1+.
4385 Description: statfs data retrieved from /proc may be stale or zero
4386 Details : When reading per-device statfs data from /proc, in the
4387 {kbytes,files}_{total,free,avail} files, it may appear
4388 as zero or be out of date.
4391 Frequency : systems with MD RAID1 external journal devices
4393 Description: lconf's call to blkid is confused by RAID1 journal devices
4394 Details : Use the "blkid -l" flag to locate the MD RAID device instead
4395 of returning all block devices that match the journal UUID.
4398 Frequency : always, for aggregate stripe size over 4GB
4400 Description: assertion fails when trying to use 4GB stripe size
4401 Details : Use "setstripe" to set stripe size over 4GB will fail the kernel,
4402 complaining "ASSERTION(lsm->lsm_xfersize != 0)"
4405 Frequency : always on ppc64
4407 Description: the first write on an ext3 filesystem with mballoc got stuck
4408 Details : ext3_mb_generate_buddy() uses find_next_bit() which does not
4409 perform endianness conversion.
4411 ------------------------------------------------------------------------------
4413 02-14-2006 Cluster File Systems, Inc. <info@clusterfs.com>
4415 * WIRE PROTOCOL CHANGE. This version of Lustre networking WILL NOT
4416 INTEROPERATE with older versions automatically. Please read the
4417 user documentation before upgrading any part of a live system.
4418 * WARNING: Lustre networking configuration changes are required with
4419 this release. See https://bugzilla.clusterfs.com/show_bug.cgi?id=10052
4422 * Support for kernels:
4423 2.6.9-22.0.2.EL (RHEL 4)
4424 2.6.5-7.244 (SLES 9)
4425 2.6.12.6 vanilla (kernel.org)
4428 Severity : enhancement
4429 Bugzilla : 7981/8208
4430 Description: Introduced Lustre Networking (LNET)
4431 Details : LNET is new networking infrastructure for Lustre, it includes
4432 a reorganized network configuration mode (see the user
4433 documentation for full details) as well as support for routing
4434 between different network fabrics. Lustre Networking Devices
4435 (LNDs) for the supported network fabrics have also been
4436 created for this new infrastructure.
4438 Severity : enhancement
4439 Description: Introduced Access control lists
4440 Details : clients can set ACLs on files and directories in order to have
4441 more fine-grained permissions than the standard Unix UGO+RWX.
4442 The MDS must be started with the "-o acl" mount option.
4444 Severity : enhancement
4445 Description: Introduced filesystem quotas
4446 Details : Administrators may now establish per-user quotas on the
4449 Severity : enhancement
4451 Description: Configuration change for the XT3
4452 The PTLLND is now used to run Lustre over Portals on the XT3
4453 The configure option(s) --with-cray-portals are no longer used.
4454 Rather --with-portals=<path-to-portals-includes> is used to
4455 enable building on the XT3. In addition to enable XT3 specific
4456 features the option --enable-cray-xt3 must be used.
4461 Description: Running on many-way SMP OSTs can trigger oops in llcd_send()
4462 Details : A race between allocating a new llcd and re-getting the llcd_lock
4463 allowed another thread to grab newly-allocated llcd.
4465 Severity : enhancement
4467 Description: 2.6 OST async journal commit and locking fix to improve performance
4468 Details : The filter_direct_io()+filter_commitrw_write() journal commits for
4469 2.6 kernels are now async as they already were in 2.4 kernels so
4470 that they can commit concurrently with the network bulk transfer.
4471 For block-allocated files the filter allocation semaphore is held
4472 to avoid filesystem fragmentation during allocation. BKL lock
4473 removed for 2.6 xattr operations where it is no longer needed.
4478 Description: lconf incorrectly determined whether two IP networks could talk
4479 Details : In some more complicated routing and multiple-network
4480 configurations, lconf will avoid trying to make a network
4481 connection to a disjoint part of the IP space. It was doing the
4482 math incorrectly for one set of cases.
4487 Description: Fix for potential infinite loop processing records in an llog.
4488 Details : If an llog record is corrupted/zeroed, it is possible to loop
4489 forever in llog_process(). Validate the llog record length
4490 and skip the remainder of the block on error.
4493 Frequency : occasional (liblustre only)
4495 Description: liblustre could not open files whose last component is a symlink
4496 Details : sysio_path_walk() would incorrectly pass the open intent to
4497 intermediate path components.
4500 Frequency : rare (liblustre only with non-standard tuning)
4501 Bugzilla : 7201 (7350)
4502 Description: Tuning the MDC DLM LRU size to zero triggers client LASSERT
4503 Details : llu_lookup_finish_locks() tries to set lock data on a lock
4504 after it has been released, only do this for referenced locks
4506 Severity : enhancement
4508 Description: specifying an (invalid) directory default stripe_size of -1
4509 would reset the directory default striping
4510 Details : stripe_size -1 was used internally to signal directory stripe
4511 removal, now use "all default" to signal dir stripe removal
4512 as a directory striping of "all default" is not useful
4515 Frequency : common for large clusters running liblustre clients
4517 Description: doing an ls when liblustre clients are running is slow
4518 Details : sending a glimpse AST to a liblustre client waits for every AST
4519 to time out, as liblustre clients will not respond. Since they
4520 cannot cache data we refresh the OST lock LVB from disk instead.
4522 Severity : enhancement
4524 Description: doing an ls at the same time as file IO can be slow
4525 Details : enqueue and other "small" requests can be blocked behind many
4526 large IO requests. Create a new OST IO portal for non-IO
4527 requests so they can be processed faster.
4530 Frequency : rare (only HPUX clients mounting unsupported re-exported NFS vol)
4532 Description: an HPUX NFS client would get -EACCESS when ftruncate()ing a newly
4533 created file with mode 000
4534 Details : the Linux NFS server relies on an MDS_OPEN_OWNEROVERRIDE hack to
4535 allow an ftruncate() as a non-root user to a file with mode 000.
4536 Lustre now respects this flag to disable mode checks when
4537 truncating a file owned by the user
4540 Frequency : liblustre-only, when liblustre client dies unexpectedly or becomes
4543 Description: Revoking locks from clients that went dead or catatonic might take
4545 Details : New lock flags FL_CANCEL_ON_BLOCK used by liblustre makes
4546 cancellation of such locks instant on servers without waiting for
4547 any reply from clients. Clients drops these locks when cancel
4548 notification from server is received without replying.
4551 Frequency : liblustre-only, when liblustre client dies or becomes busy
4553 Description: Doing ls on Linux clients can take a long time with active
4555 Details : Liblustre client cannot handle ASTs in timely manner, so avoid
4556 granting such locks to it in the first place if possible. Locks
4557 are taken by proxy on the OST during the read or write and
4558 dropped immediately afterward. Add connect flags handling, do
4559 not grant locks to liblustre clients for glimpse ASTs.
4561 Severity : enhancement
4563 Description: Improve read-ahead algorithm to avoid excessive IO for random reads
4564 Details : Existing read-ahead algorithm is tuned for the case of streamlined
4565 sequential reads and behaves badly with applications doing random
4566 reads. Improve it by reading ahead at least read region, and
4567 avoiding excessive large RPC for small reads.
4569 Severity : enhancement
4571 Description: Creating more than 1000 files for a single job may cause a load
4572 imbalance on the OSTs if there are also a large number of OSTs.
4573 Details : qos_prep_create() uses an OST index reseed value that is an
4574 even multiple of the number of available OSTs so that if the
4575 reseed happens in the middle of the object allocation it will
4576 still utilize the OSTs as uniformly as possible.
4581 Description: OST or MDS may oops in ping_evictor_main()
4582 Details : ping_evictor_main() drops obd_dev_lock if deleting a stale export
4583 but doesn't restart at beginning of obd_exports_timed list
4586 Severity : enhancement
4588 Description: improve by-nid export eviction on the MDS and OST
4589 Details : allow multiple exports with the same NID to be evicted at one
4590 time without re-searching the exports list.
4593 Frequency : rare, only with supplementary groups enabled on SMP 2.6 kernels
4595 Description: MDS may oops in groups_free()
4596 Details : in rare race conditions a newly allocated group_info struct is
4597 freed again, and this can be NULL. The 2.4 compatibility code
4598 for groups_free() checked for a NULL pointer, but 2.6 did not.
4601 Frequency : common for liblustre clients doing little filesystem IO
4602 Bugzilla : 9352, 7313
4603 Description: server may evict liblustre clients accessing contended locks
4604 Details : if a client is granted a lock or receives a completion AST
4605 with a blocking AST already set it would not reply to the AST
4606 for LDLM_FL_CANCEL_ON_BLOCK locks. It now replies to such ASTs.
4609 Frequency : lfs setstripe, only systems with more than 160 OSTs
4611 Description: unable to set striping with a starting offset beyond OST 160
4612 Details : llapi_create_file() incorrectly limited the starting stripe
4613 index to the maximum single-file stripe count.
4616 Frequency : LDAP users only
4618 Description: lconf did not handle in-kernel recovery with LDAP properly
4619 Details : lconf/LustreDB get_refs() is searching the wrong namespace
4621 Severity : enhancement
4623 Description: bind OST threads to NUMA nodes to improve performance
4624 Details : all OST threads are uniformly bound to CPUs on a single NUMA
4625 node and do their allocations there to localize memory access
4627 Severity : enhancement
4629 Description: llmount can determine client NID directly from Myrinet (GM)
4630 Details : the client NID code from gmnalnid was moved directly into
4631 llmount, removing the need to use this or specifying the
4632 client NID explicitly when mounting GM clients with zeroconf
4635 Frequency : if client is started with down MDS
4637 Description: if client is started with down MDS mount hangs in ptlrpc_queue_wait
4638 Details : Having an LWI_INTR() wait event (interruptible, but no timeout)
4639 will wait indefinitely in ptlrpc_queue_wait->l_wait_event() after
4640 ptlrpc_import_delayed_req() because we didn't check if the
4641 request was interrupted, and we also didn't break out of the
4642 event loop if there was no timeout
4647 Description: data loss during non-page-aligned writes to a single file from
4648 both multiple nodes and multiple threads on one node at same time
4649 Details : updates to KMS and lsm weren't protected by common lock. Resulting
4650 inconsistency led to false short-reads, that were cached and later
4651 used by ->prepare_write() to fill in partially written page,
4652 leading to data loss.
4655 Frequency : always, if lconf --abort_recovery used
4657 Description: lconf --abort_recovery fails with 'Operation not supported'
4658 Details : lconf was attempting to abort recovery on the MDT device and not
4661 Severity : enhancement
4663 Description: remove cleanup logs
4664 Details : replace lconf-generated cleanup logs with lustre internal
4665 cleanup routines. Eliminates the need for client-cleanup and
4668 Severity : enhancement
4670 Description: add support for EAs (user and system) on lustre filesystems
4671 Details : it is now possible to store extended attributes in the Lustre
4672 client filesystem, and with the user_xattr mount option it
4673 is possible to allow users to store EAs on their files also
4675 Severity : enhancement
4677 Description: Add possibility (config option) to show minimal available OST free
4679 Details : When compiled with --enable-mindf configure option, statfs(2)
4680 (and so, df) will return least minimal free space available from
4681 all OSTs as amount of free space on FS, instead of summary of
4682 free spaces of all OSTs.
4684 Severity : enhancement
4686 Description: do not expand extent locks acquired on OST-side
4687 Details : Modify ldlm_extent_policy() to not expand local locks, acquired
4688 by server: they are not cached anyway.
4691 Frequency : when mmap is used/binaries executed from Lustre
4693 Description: Unmmap pages before throwing them away from read cache.
4694 Details : llap_shrink cache now attempts to unmap pages before discarding
4695 them (if unmapping failed - do not discard). SLES9 kernel has
4696 extra checks that trigger if this unmapping is not done first.
4701 Description: lconf didn't resolve symlinks before checking to see whether a
4702 given mountpoint was already in use
4705 Frequency : when migrating failover services
4706 Bugzilla : 6395, 9514
4707 Description: When migrating a subset of services from a node (e.g. failback
4708 from a failover service node) the remaining services would
4709 time out and evict clients.
4710 Details : lconf --force (implied by --failover) sets the global obd_timeout
4711 to 5 seconds in order to quickly disconnect, but this caused
4712 other RPCs to time out too quickly. Do not change the global
4713 obd_timeout for force cleanup, only set it for DISCONNECT RPCs.
4715 Severity : enhancement
4716 Frequency : if MDS is started with down OST
4717 Bugzilla : 9439,5706
4718 Description: Allow startup/shutdown of an MDS without depending on the
4719 availability of the OSTs.
4720 Details : Asynchronously call mds_lov_synchronize during MDS startup.
4721 Add appropriate locking and lov-osc refcounts for safe
4722 cleaning. Add osc abort_inflight calls in case the
4723 synchronize never started.
4726 Frequency : occasional (Cray XT3 only)
4728 Description: root not authorized to access files in CRAY_PORTALS environment
4729 Details : The client process capabilities were not honoured on the MDS in
4730 a CRAY_PORTALS/CRAY_XT3 environment. If the file had previously
4731 been accessed by an authorized user then root was able to access
4732 the file on the local client also. The root user capabilities
4733 are now allowed on the MDS, as this environment has secure UID.
4736 Frequency : occasional
4738 Description: ldiskfs "too long searching" message happens too often
4739 Details : A debugging message (otherwise harmless) prints too often on
4740 the OST console. This has been reduced to only happen when
4741 there are fragmentation problems on the filesystem.
4746 Description: Division by zero in statfs when all OSCs are inactive
4747 Details : lov_get_stripecnt() returns zero due to incorrect order of checks,
4748 lov_statfs divides by value returned by lov_get_stripecnt().
4752 Bugzilla : 9489, 3273
4753 Description: First write from each client to each OST was only 4kB in size,
4754 to initialize client writeback cache, which caused sub-optimal
4755 RPCs and poor layout on disk for the first written file.
4756 Details : Clients now request an initial cache grant at (re)connect time
4757 and so that they can start streaming writes to the cache right
4758 away and always do full-sized RPCs if there is enough data.
4759 If the OST is rebooted the client also re-establishes its grant
4760 so that client cached writes will be honoured under the grant.
4765 Description: Slow ls (and stat(2) syscall) on files residing on IO-loaded OSTs
4766 Details : Now I/O RPCs go to different portal number and (presumably) fast
4767 lock requests (and glimses) and other RPCs get their own service
4768 threads pool that should be able to service those RPCs
4771 Severity : enhancement
4773 Description: Ability to exchange lustre version between client and servers and
4774 issue warnings at client side if client is too old. Also for
4775 liblustre clients there is ability to refuse connection of too old
4777 Details : New 'version' field is added to connect data structure that is
4778 filled with version info. That info is later checked by server and
4782 Frequency : rare, liblustre only.
4783 Bugzilla : 9296, 9581
4784 Description: Two simultaneous writes from liblustre at offset within same page
4785 might proceed at the same time overwriting eachother with stale
4787 Details : I/O lock withing llu_file_prwv was released too early, before data
4788 actually was hitting the wire. Extended lock-holding time until
4789 server acknowledges receiving data.
4792 Frequency : extremely rare. Never observed in practice.
4794 Description: avoid generating lustre_handle cookie of 0.
4795 Details : class_handle_hash() generates handle cookies by incrementing
4796 global counter, and can hit 0 occasionaly (this is unlikely, but
4797 not impossible, because initial value of cookie counter is
4798 selected randonly). Value of 0 is used as a sentinel meaning
4799 "unassigned handle" --- avoid it. Also coalesce two critical
4800 sections in this function into one.
4802 Severity : enhancement
4804 Description: allow liblustre clients to delegate truncate locking to OST
4805 Details : To avoid overhead of locking, liblustre client instructs OST to
4806 take extent lock in ost_punch() on client's behalf. New connection
4807 flag is added to handle backward compatibility.
4809 Severity : enhancement
4810 Bugzilla : 4928, 7341, 9758
4811 Description: allow number of OST service threads to be specified
4812 Details : a module parameter allows the number of OST service threads
4813 to be specified via "options ost ost_num_threads={N}" in the
4814 OSS's /etc/modules.conf or /etc/modprobe.conf.
4818 Bugzilla : 6146, 9635, 9895
4819 Description: servers crash with bad pointer in target_handle_connect()
4820 Details : In rare cases when a client is reconnecting it was possible that
4821 the connection request was the last reference for that export.
4822 We would temporarily drop the export reference and get a new
4823 one, but this may have been the last reference and the export
4824 was just destroyed. Get new reference before dropping old one.
4826 Severity : enhancement
4827 Frequency : if client is started with failover MDS
4829 Description: Allow multiple MDS hostnames in the mount command
4830 Details : Try to read the configuration from all specified MDS
4831 hostnames during a client mount in case the "primary"
4834 Severity : enhancement
4836 Description: Stop sending data to evicted clients as soon as possible.
4837 Details : Check if the client we are about to send or are sending data to
4838 was evicted already. (Check is done every second of waiting,
4839 for which l_wait_event interface was extended to allow checking
4840 of exit condition at specified intervals).
4843 Frequency : rare, normally only when NFS exporting is done from client
4845 Description: 'bad disk LOV MAGIC: 0x00000000' error when chown'ing files
4847 Details : Make mds_get_md() recognise empty md case and set lmm size to 0.
4850 Frequency : always, if srand() is called before liblustre initialization
4852 Description: Liblustre uses system PRNG disturbing its usage by user application
4853 Details : Introduce internal to lustre fast and high-quality PRNG for
4854 lustre usage and make liblustre and some other places in generic
4855 lustre code to use it.
4857 Severity : enhancement
4858 Bugzilla : 9477, 9557, 9870
4859 Description: Verify that the MDS configuration logs are updated when xml is
4860 Details : Check if the .xml configuration logs are newer than the config
4861 logs stored on the MDS and report an error if this is the case.
4862 Request --write-conf, or allow starting with --old_conf.
4864 Severity : enhancement
4866 Description: Handle symlinks in the path when checking if Lustre is mounted.
4867 Details : Resolve intermediate symlinks when checking if a client has
4868 mounted a filesystem to avoid duplicate client mounts.
4873 Description: lconf can hit an error exception but still return success.
4874 Details : The lconf command catches the Command error exception at the top
4875 level script context and will exit with the associated exit
4876 status, but doesn't ensure that this exit status is non-zero.
4881 Description: failure of ptlrpc thread startup can cause oops
4882 Details : Starting a ptlrpc service thread can fail if there are a large
4883 number of threads or the server memory is very fragmented.
4884 Handle this without oopsing.
4887 Frequency : always, only if liblustre and non-default acceptor port was used
4889 Description: liblustre cannot connect to servers with non-default acceptor port
4890 Details : tcpnal_set_default_params() was not called and was therefore
4891 ignoring the environment varaible TCPNAL_PORT, as well as other
4892 TCPNAL_ environment variables
4897 Description: two objects could be created on the same OST for a single file
4898 Details : If an OST is down, in some cases it was possible to create two
4899 objects on a single OST for a single file. No problems other
4900 than potential performance impact and spurious error messages.
4904 Bugzilla : 5681, 9562
4905 Description: Client may oops in ll_unhash_aliases
4906 Details : Client dcache may become inconsistent in race condition.
4907 In some cases "getcwd" can fail if the current directory is
4913 Description: Inode refcounting problems in NFS export code
4914 Details : link_raw functions used to call d_instantiate without obtaining
4915 extra inode reference first.
4919 Bugzilla : 9942, 9903
4920 Description: Referencing freed requests leading to crash, memleaks with NFS.
4921 Details : We used to require that call to ll_revalidate_it was always
4922 followed by ll_lookup_it. Also with revalidate_special() it is
4923 possible to call ll_revalidate_it() twice for the same dentry
4924 even if first occurence returned success. This fix changes semantic
4925 between DISP_ENQ_COMPLETE disposition flag to mean there is extra
4926 reference on a request referred from the intent.
4927 ll_intent_release() then releases such a request.
4930 Frequency : rare, normally benchmark loads only
4932 Description: unlinked inodes were kept in memory on the client
4933 Details : If a client is repeatedly creating and unlinking files it
4934 can accumulate a lot of stale inodes in the inode slab cache.
4935 If there is no other client load running this can cause the
4936 client node to run out of memory. Instead flush old inodes
4937 from client cache that have the same inode number as a new inode.
4940 Frequency : SLES9 2.6.5 kernel and long filenames only
4941 Bugzilla : 9969, 10379
4942 Description: utime reports stale NFS file handle
4943 Details : SLES9 uses out-of-dentry names in some cases, which confused
4944 the lustre dentry revalidation. Change it to always use the
4948 Frequency : rare, unless heavy write-truncate concurrency is continuous
4949 Bugzilla : 4180, 6984, 7171, 9963, 9331
4950 Description: OST becomes very slow and/or deadlocked during object unlink
4951 Details : filter_destroy() was holding onto the parent directory lock
4952 while truncating+unlinking objects. For very large objects this
4953 may block other threads for a long time and slow overall OST
4954 responsiveness. It may also be possible to get a lock ordering
4955 deadlock in this case, or run out of journal credits because of
4956 the combined truncate+unlink. Solution is to do object truncate
4957 first in one transaction without parent lock, and then do the
4958 final unlink in a new transaction with the parent lock. This
4959 reduces the lock hold time dramatically.
4962 Frequency : rare, 2.4 kernels only
4964 Description: MDS or OST cleanup may trip kernel BUG when dropping kernel lock
4965 Details : mds_cleanup() and filter_cleanup() need to drop the kernel lock
4966 before unmounting their filesystem in order to avoid deadlock.
4967 The kernel_locked() function in 2.4 kernels only checks whether
4968 the kernel lock is held, not whether it is this process that is
4969 holding it as 2.6 kernels do.
4974 Description: MDS or OST may oops/LBUG if a client is connecting multiple times
4975 Details : The client ptlrpc code may be trying to reconnect to a down
4976 server before a previous connection attempt has timed out.
4977 Increase the reconnect interval to be longer than the connection
4978 timeout interval to avoid sending duplicate connections to
4982 Frequency : echo_client brw_test command
4984 Description: fix echo_client to work with OST preallocated code
4985 Details : OST preallocation code (5137) didn't take echo_client IO path
4986 into account: echo_client calls filter methods outside of any
4987 OST thread and, hence, there is no per-thread preallocated
4988 pages and buffers to use. Solution: hijack pga pages for IO. As
4989 a byproduct, this avoids unnecessary data copying.
4993 Bugzilla : 3555, 5962, 6025, 6155, 6296, 9574
4994 Description: Client can oops in mdc_commit_close() after open replay
4995 Details : It was possible for the MDS to return an open request with no
4996 transaction number in mds_finish_transno() if the client was
4997 evicted, but without actually returning an error. Clients
4998 would later try to replay that open and may trip an assertion
4999 Simplify the client close codepath, and always return an error
5000 from the MDS in case the open is not successful.
5003 Frequency : rare, 2.6 OSTs only
5005 Description: OST may deadlock under high load on fragmented files
5006 Details : If there was a heavy load and highly-fragmented OST filesystems
5007 it was possible to have all the OST threads deadlock waiting on
5008 allocation of biovecs, because the biovecs were not released
5009 until the entire RPC IO was completed. Instead, release biovecs
5010 as soon as they are complete to ensure forward IO progress.
5012 Severity : enhancement
5014 Description: Support for specifying external journal device at mount
5015 Details : If an OST or MDS device is formatted with an external journal
5016 device, this device major/minor is stored in the ext3 superblock
5017 and may not be valid for failover. Allow detecting and
5018 specifying the external journal at mount time.
5023 Description: Mounting an MDS with pending unlinked files may cause oops
5024 Details : target_finish_recovery() calls mds_postrecov() which returned
5025 the number of orphans unlinked. mds_lov_connect->mds_postsetup()
5026 considers this an error and immediately begins cleaning up the
5027 lov, just after starting the mds_lov process
5029 Severity : enhancement
5031 Description: Implement 'lfs df' to report actual free space on per-OST basis
5032 Details : Add sub-command 'df' on 'lfs' to report the disk space usage of
5033 MDS/OSDs. Usage: lfs df [-i][-h]. Command Options: '-i' to report
5034 usage of objects; '-h' to report in human readable format.
5036 ------------------------------------------------------------------------------
5038 08-26-2005 Cluster File Systems, Inc. <info@clusterfs.com>
5045 Description: Mounting an ldiskfs file system with mballoc may crash OST node.
5046 Details : ldiskfs mballoc code may reference an uninitialized buddy struct
5047 at startup during orphan unlinking. Instead, skip buddy update
5048 before setup, as it will be regenerated after recovery is complete.
5053 Description: If an OST is inactive, its locks might reference stale inodes.
5054 Details : lov_change_cbdata() must iterate over all namespaces, even if
5055 they are inactive to clear inode references from the lock.
5057 Severity : enhancement
5058 Frequency : occasional, if non-standard max_dirty_mb used
5060 Description: Client will block write RPCs if not enough grant
5061 Details : If a client has max_dirty_mb smaller than max_rpcs_in_flight,
5062 then the client will block writes while waiting for another RPC
5063 to complete instead of consuming its dirty limit. With change
5064 we get improved performance when max_dirty_mb is small.
5066 Severity : enhancement
5067 Bugzilla : 3389, 6253
5068 Description: Add support for supplementary groups on the MDS.
5069 Details : The MDS has an upcall /proc/fs/lustre/mds/{mds}/group_upcall
5070 (set to /usr/sbin/l_getgroups if enabled) which will do MDS-side
5071 lookups for user supplementary groups into a cache.
5075 Description: O_CREAT|O_EXCL open flags in liblustre always return -EEXIST
5076 Details : Make libsysio to not enforce O_EXCL by clearing the flag,
5077 for liblustre O_EXCL is enforced by MDS.
5081 Description: readdir never returns NULL in liblustre.
5082 Details : Corrected llu_iop_getdirentries logic, to return offset of next
5083 dentry in struct dirent.
5087 Frequency : liblustre only, depends on application IO pattern
5088 Description: liblustre clients evicted if not contacting servers
5089 Details : Don't put liblustre clients into the ping_evictor list, so
5090 they will not be evicted by the pinger ever.
5092 Severity : enhancement
5094 Description: Add ability to evict clients by NID from MDS.
5095 Details : By echoing "nid:$NID" string into
5096 /proc/fs/lustre/mds/.../evict_client client with nid that equals to
5097 $NID would be instantly evicted from this MDS and from all active
5098 OSTs connected to it.
5102 Description: Do not query file size twice, somewhat slowing stat(2) calls.
5103 Details : lookup_it_finish() used to query file size from OSTs that was not
5108 Description: service threads change working directory to that of init
5109 Details : Starting lustre service threads may pin the working directory
5110 of the parent thread, making that filesystem busy. Threads
5111 now change to the working directory of init to avoid this.
5115 Frequency : during shutdown only
5116 Description: shutdown with a failed MDS or OST can cause unmount to hang
5117 Details : Don't resend DISCONNECT messages in ptlrpc_disconnect_import()
5122 Frequency : 2.6 only
5123 Description: chmod/chown may include an extra supplementary group
5124 Details : ll{,u}_mdc_pack_op_data() does not properly initialize the
5125 supplementary group and if none is specified this is used.
5128 Bugzilla : 5479 (6816)
5130 Description: Racing open + rm can assert client in mdc_set_open_replay_data()
5131 Details : If lookup is in progress on a file that is unlinked we might try
5132 to revalidate the inode and fail in revalidate after lookup is
5133 complete and ll_file_open() enqueues the open again but
5134 it_open_error() was not checking DISP_OPEN_OPEN errors correctly.
5137 Frequency : always, if lconf --abort_recovery used
5139 Description: lconf --abort_recovery fails with 'Operation not supported'
5140 Details : lconf was attempting to abort recovery on the MDT device and not
5143 ------------------------------------------------------------------------------
5145 2005-08-08 Cluster File Systems, Inc. <info@clusterfs.com>
5150 Frequency : rare (only unsupported configurations with a node running as an
5152 Bugzilla : 6514, 5137
5153 Description: Mounting a Lustre file system on a node running as an OST could
5155 Details : OSTs now preallocates memory needed to write out data at
5156 startup, instead of when needed, to avoid having to
5157 allocate memory in possibly low memory situations.
5158 Specifically, if the file system is mounted on on OST,
5159 memory pressure could force it to try to write out data,
5160 which it needed to allocate memory to do. Due to the low
5161 memory, it would be unable to do so and the node would
5162 become unresponsive.
5164 Severity : enhancement
5166 Description: Addition of lconf --service command line option
5167 Details : lconf now accepts a '--service <arg>' option, which is
5168 shorthand for 'lconf --group <arg> --select <arg>=<hostname>'
5170 Severity : enhancement
5172 Description: Failover mode is now the default for OSTs.
5173 Details : By default, OSTs will now run in failover mode. To return to
5174 the old behaviour, add '--failout' to the lmc line for OSTs.
5176 Severity : enhancement
5178 Description: Health checks are now provided for MDS and OSTs
5179 Details : Additional detailed health check information on MSD and OSTs
5180 is now provided through the procfs health_check value.
5183 Frequency : occasional, depends on IO load
5185 Description: Disk fragmentation on the OSTs could eventually cause slowdowns
5186 after numerous create/delete cycles
5187 Details : The ext3 inode allocation policy would not allocate new inodes
5188 very well on the OSTs because there are no new directories
5189 being created. Instead we look for groups with free space if
5190 the parent directories are nearly full.
5195 Description: Network or server problems during mount may cause partially
5196 mounted clients instead of returning an error.
5197 Details : The config llog parsing code may overwrite the error return
5198 code during mount error handling, returning success instead
5204 Description: MDS can fail to allocate large reply buffers
5205 Details : After long uptimes the MDS can fail to allocate large reply
5206 buffers (e.g. zconf client mount config records) due to memory
5207 fragmentation or consumption by the buffer cache. Preallocate
5208 some large reply buffers so that these replies can be sent even
5209 under memory pressure.
5213 Frequency : rare (liblustre)
5214 Description: fsx running with liblustre complained that using truncate() to
5215 extend the file doesn't work. This patch corrects that issue.
5216 Details : This is the liblustre equivalent of the fix for bug 6196. Fixes
5217 ATTR_SIZE and lsm use in llu_setattr_raw.
5221 Frequency : rare, only 2.6 kernels
5222 Description: Unusual file access patterns on the MDS may result in inode
5223 data being lost in very rare circumstances.
5224 Details : Bad interaction between the ea-in-inode patch and the "no-read"
5225 code in the 2.6 kernel caused the inode and/or EA data not to
5226 be read from disk, causing single-file corruption.
5230 Frequency : rare, only 2.6 filesystems using extents
5231 Description: Heavy concurrent write and delete load may cause data corruption.
5232 Details : It was possible under high-load situations to have an extent
5233 metadata block in the block device cache from a just-unlinked
5234 file overwrite a newly-allocated data block. We now unmap any
5235 metadata buffers that alias just-allocated data blocks.
5239 Frequency : filesystems with default stripe_count larger than 77
5240 Description: lconf+mke2fs fail when formatting filesystem with > 77 stripes
5241 Details : lconf specifies an inode size of 4096 bytes when the default
5242 stripe_count is larger than 77. This conflicts with the default
5243 inode density of 1 per 4096 bytes. Allocate smaller inodes in
5244 this case to avoid pinning too much memory for large EAs.
5246 ------------------------------------------------------------------------------
5248 2005-07-07 Cluster File Systems, Inc. <info@clusterfs.com>
5253 Frequency : rare (extremely heavy IO load with hundreds of clients)
5255 Description: Client is evicted, gets IO error writing to file
5256 Details : lock ordering changes for bug 5492 reintroduced bug 3267 and
5257 caused clients to be evicted for AST timeouts. The fixes in
5258 bug 5192 mean we no longer need to have such short AST timeouts
5259 so ldlm_timeout has been increased.
5262 Frequency : occasional during --force or --failover shutdown under load
5263 Bugzilla : 5949, 4834
5264 Description: Server oops/LBUG if stopped with --force or --failover under load
5265 Details : a collection of import/export refcount and cleanup ordering
5266 issues fixed for safer force cleanup
5269 Frequency : only filesystems larger than 120 OSTs
5270 Bugzilla : 5990, 6223
5271 Description: lfs getstripe would oops on a very large filesystem
5272 Details : lov_getconfig used kfree on vmalloc'd memory
5275 Frequency : only filesystems exporting via NFS to Solaris 10 clients
5276 Bugzilla : 6242, 6243
5277 Description: reading from files that had been truncated to a non-zero size
5278 but never opened returned no data
5279 Details : ll_file_read() reads zeros from no-object files to EOF
5284 Description: A bug in MDS/OSS recovery could cause the OSS to fail an assertion
5285 Details : There's little harm in aborting MDS/OSS recovery and letting it
5286 try again, so I removed the LASSERT and return an error instead.
5288 Severity : enhancement
5290 Description: New debugging infrastructure for tracking down data corruption
5291 Details : The I/O checksum code was replaced to: (a) control it at runtime,
5292 (b) cover more of the client-side code path, and (c) try to narrow
5293 down where problems occurred
5297 Bugzilla : 3819, 4364, 4397, 6313
5298 Description: Racing close and eviction MDS could cause assertion in mds_close
5299 Details : It was possible to get multiple mfd references during close and
5300 client eviction, leading to one thread referencing a freed mfd.
5302 Severity: : enhancement
5303 Bugzilla : 3262, 6359
5304 Description: Attempts to reconnect to servers are now more aggressive.
5305 Details : This builds on the enhanced upcall-less recovery that was added
5306 in 1.4.2. When trying to reconnect to servers, clients will
5307 now try each server in the failover group every 10 seconds. By
5308 default, clients would previously try one server every 25 seconds.
5313 Description: After recovery, certain operations trigger a failed
5314 assertion on a client.
5315 Details : Failing over an mds, using lconf -d --failover, while a
5316 client was doing a readdir() call would cause the client to
5317 LBUG after recovery completed and the readdir() was resent.
5319 Severity : enhancement
5321 Description: Default groups are now added by lconf
5322 Details : You can now run lconf --group <servicename> without having to
5323 manually add groups with lmc.
5326 Frequency : occasional
5328 Description: Nodes with an elan id of 0 trigger a failed assertion
5331 Frequency : always when accessing e.g. tty/console device nodes
5333 Description: tty and some other devices nodes cannot be used on lustre
5334 Details : file's private_data field is used by device data and lustre
5335 values in there got lost. New field was added to struct file to
5336 store fs-specific private data.
5339 Frequency : when exporting Lustre via NFS
5341 Description: NFSD failed occasionally when looking up a path component
5342 Details : NFSD is looking up ".." which was broken in ext3 directories
5343 that had grown large enough to become hashed.
5346 Frequency : Clusters with multiple interfaces not on the same subnet
5348 Description: Nodes will repeatedly try to reconnect to an interface which it
5349 cannot reach and report an error to the log.
5350 Details : Extra peer list entries will be created by lconf with some peers
5351 unreachable. lconf now validates the peer before adding it.
5354 Frequency : Only if a default stripe is set on the filesystem root.
5356 Description: Setting a default stripe on the filesystem root prevented the
5357 filesystem from being remounted.
5358 Details : The client was sending extra request flags in the root getattr
5359 request and did not allocate a reply buffer for the dir EA.
5362 Frequency : occasional, higher if lots of files are accessed by one client
5363 Bugzilla : 6159, 6097
5364 Description: Client trips assertion regarding lsm mismatch/magic
5365 Details : While revalidating inodes the VFS looks up inodes with ifind()
5366 and in rare cases can find an inode that is being freed.
5367 The ll_test_inode() code will free the lsm during ifind()
5368 when it finds an existing inode and then the VFS later attaches
5369 this free lsm to a new inode.
5373 Bugzilla : 6422, 7030
5374 Description: MDS deadlock between mkdir and client eviction
5375 Details : Creating a new file via mkdir or mknod (starting a transaction
5376 and getting the ns lock) can deadlock with client eviction
5377 (gets ns lock and trying to finish a synchronous transaction).
5380 Frequency : occasional
5381 Description: While starting a server, the fsfilt_ext3 module could not be
5383 Details : CFS's improved ext3 filesystem is named ldiskfs for 2.6
5384 kernels. Previously, lconf would still use the ext3 name
5385 when trying to load modules. Now, it will correctly use
5386 ext3 on 2.4 and ldiskfs on 2.6.
5388 Severity : enhancement
5389 Description: The default stripe count has been changed to 1
5390 Details : The interpretation of the default stripe count (0, to lfs
5391 or lmc) has been changed to mean striping across a single
5392 OST, rather than all available. For general usage we have
5393 found a stripe count of 1 or 2 works best.
5395 Severity : enhancement
5396 Description: Add support for compiling against Cray portals.
5397 Details : Conditional compiling for some areas that are different
5401 Frequency : occasional
5402 Bugzilla : 6409, 6834
5403 Description: Creating files with an explicit stripe count may lead to
5404 a failed assertion on the MDS
5405 Details : If some OSTs are full or unavailable, creating files may
5406 trigger a failed assertion on the MDS. Now, Lustre will
5407 try to use other servers or return an error to the
5411 Frequency : occasional
5413 Description: Multiple concurrent overlapping read+write on multiple SMP nodes
5414 caused lock timeout during readahead (since 1.4.2).
5415 Details : Processes doing readahead might match a lock that hasn't been
5416 granted yet if there are overlapping and conflicting lock
5417 requests. The readahead process waits on ungranted lock
5418 (original lock is CBPENDING), while OST waits for that process
5419 to cancel CBPENDING read lock and eventually evicts client.
5421 Severity : enhancement
5423 Description: Initial enabling of flock support for clients
5424 Details : Implements fcntl advisory locking and file status functions.
5425 This feature is provided as an optional mount flag (default
5426 off), and is NOT CURRENTLY SUPPORTED. Not all types of record
5427 locking are implemented yet, and those that are are not guaranteed
5428 to be completely correct in production environments.
5429 mount -t lustre -o [flock|noflock] ...
5432 Frequency : occasional
5434 Description: OSTs running 2.4 kernels but with extents enabled might trip an
5435 assertion in the ext3 JBD (journaling) layer.
5436 Details : The b_committed_data struct is protected by the big kernel lock
5437 in 2.4 kernels, serializing journal_commit_transaction() and
5438 ext3_get_block_handle->ext3_new_block->find_next_usable_block()
5439 access to this struct. In 2.6 kernels there is finer grained
5440 locking to improve SMP performance of the JBD layer.
5444 Description: Changes the "SCSI I/O Stats" kernel patch to default to "enabled"
5449 Description: merge and cleanup kernel patches.
5452 -----------------------------------------------------------------------------
5454 2005-05-05 Cluster File Systems, Inc. <info@clusterfs.com>
5456 NOTE: Lustre 1.4.2 uses an incompatible network protocol than previous
5457 versions of Lustre. Please update all servers and clients to
5458 version 1.4.2 or later at the same time. You must also run
5459 "lconf --write-conf {config}.xml" on the MDS while it is stopped
5460 to update the configuration logs.
5462 - fix for HPUX NFS client breakage when NFS exporting Lustre (5781)
5463 - mdc_enqueue does not need max_mds_easize request buffer on send (5707)
5464 - swab llog records of type '0' so we get proper header size/idx (5861)
5465 - send llog cancel req to DLM cancel portal instead of cb portal (5515)
5466 - fix rename of one directory over another leaking an inode (5953)
5467 - avoid SetPageDirty on 2.6 (5981)
5468 - don't re-add just-being-destroyed locks to the waiting list (5653)
5469 - when creating new directories, inherit the parent's custom
5470 striping settings if present parent (3048)
5471 - flush buffers from cache before direct IO in 2.6 obdfilter (4982)
5472 - don't hold i_size_sem in ll_nopage() and ll_ap_refresh_count (6077)
5473 - don't hold client locks on temporary worklist from l_lru (5666)
5474 - handle IO errors in 2.6 obdfilter bio completion routine (6046)
5475 - automatically evict dead clients (5921)
5476 - Update file size properly in create+truncate+fstat case (6196)
5477 - Do not unhash mountpoint dentries, do not allow removal of
5479 - Avoid lock ordering deadlock issue with write/truncate (6203,5654)
5480 - reserve enough journal credits in fsfilt_start_log for setattr (4554)
5481 - ldlm_enqueue freed-export error path would always LBUG (6149,6184)
5482 - don't reference lr_lvb_data until after we hold lr_lvb_sem (6170)
5483 - don't overwrite last_rcvd if there is a *_client_add() error (6086)
5484 - Correctly handle reads of files with no objects (6243)
5485 - lctl recover will also mark a device active if deactivate used (5933)
5487 - by default create 1 inode per 4kB space on MDS, per 16kB on OSTs
5488 - allow --write-conf on an MDS with different nettype than client (5619)
5489 - don't write config llogs to MDS for mounts not from that MDS (5617)
5490 - lconf should create multiple TCP connections from a client (5201)
5491 - init scripts are now turned off by default; run chkconfig --on
5492 lustre and chkconfig --on lustrefs to use them
5493 - upcalls are no longer needed for clients to recover to failover
5495 - add --abort-recovery option to lconf to abort recovery on device
5497 - add support for an arbitrary number of OSTs (3026)
5498 - Quota support protocol changes.
5499 - forward compatibility changes to wire structs (6007)
5500 - rmmod NALs that might be loaded because of /etc/modules.conf (6133)
5501 - support for mountfsoptions and clientoptions to the Lustre LDAP (5873)
5502 - improved "lustre status" script
5503 - initialize blocksize for non-regular files (6062)
5504 - added --disable-server and --disable-client configure options (5782)
5505 - introduce a lookup cache for lconf to avoid repeated DB scans (6204)
5506 - Vanilla 2.4.29 support
5507 - increase maximum number of obd devices to 520 (6242)
5508 - remove the tcp-zero-copy patch from the suse-2.4 series (5902)
5509 - Quadrics Elan drivers are now included for the RHEL 3 2.4.21 and
5510 SLES 9 2.6.5 kernels
5511 - limit stripes per file to 160 (the maximum EA size) (6093)
5513 2005-03-22 Cluster File Systems, Inc. <info@clusterfs.com>
5516 - don't LASSERT in ll_release on NULL lld with NFS export (4655, 5760)
5517 - hold NS lock when calling handle_ast_error->del_waiting_lock (5746)
5518 - fix setattr mtime regression from lovcleanup merge (4829, 5669)
5519 - workaround for 2.6 crash in ll_unhash_aliases (5687, 5210)
5520 - small ext3 extents cleanups and fixes (5733)
5521 - improved mballoc code, several small races and bugs fixed (5733, 5638)
5522 - kernel version 43 - fix remove_suid bugs in both 2.4 and 2.6 (5695)
5523 - avoid needless client->OST connect, fix handle mismatch (5317)
5524 - fix DLM error path that led to out-of-sync client, long delays (5779)
5525 - support common vfs-enforced mount options (nodev,nosuid,noexec) (5637)
5526 - fix several locking issues related to i_size (5492,5624,5654,5672)
5527 - don't move pending lock onto export if it is already evicted (5683)
5528 - fix kernel oops when creating .foo in unlinked directory (5548)
5529 - fix deadlock in obdfilter statistics vs. object create (5811)
5530 - use time_{before,after} to avoid timer jiffies wrap (5882)
5531 - shutdown --force/--failover stability (3607,3651,4797,5203,4834)
5532 - Do not leak request if server was not able to process it (5154)
5533 - If mds_open unable to find parent dir, make that negative lookup(5154)
5534 - don't create new directories with extent-mapping (5909, 5936)
5536 - fix lustre/lustrefs init scripts for SuSE (patch from Scali, 5702)
5537 - don't hold the pinger_sem in ptlrpc_pinger_sending_on_import
5538 - change obd_increase_kms to obd_adjust_kms (up or down) (5654)
5539 - lconf, lmc search both /usr/lib and /usr/lib64 for Python libs (5800)
5540 - support for RHEL4 kernel on i686 (5773)
5541 - provide error messages when incompatible logs are encountered (5898)
5543 2005-02-18 Cluster File Systems, Inc. <info@clusterfs.com>
5544 * version 1.4.0.10 (1.4.1 release candidate 1)
5546 - don't keep a lock reference when lock is not granted (4238)
5547 - unsafe list practices (rarely) led to infinite eviction loop (4908)
5548 - add per-fs limit of Lustre pages in page cache, avoid OOM (4699)
5549 - drop import inflight refcount on signal_completed_replay error (5255)
5550 - unlock page after async write error during send (3677)
5551 - handle missing objects in filter_preprw_read properly (5265)
5552 - no transno return for symlink open, don't save no-trasno open (3440)
5553 - don't try to complete elan receive that already failed (4012)
5554 - free RPC server reply state on error (5406)
5555 - clean up thread from ptlrpc_start_thread() on error (5160)
5556 - readahead could read extra page into cache that wasn't ejected (5388)
5557 - prevent races in class_attach/setup/cleanup/detach (5260)
5558 - don't dereference de->d_inode after l_dput of de (5458)
5559 - use "int" for stripe value returned from lock_to_stripe (5544)
5560 - mballoc allocation and error-checking fixes in 2.6 (5504)
5561 - block device patches to fix I/O request sizes in 2.6 (5482)
5562 - look up hostnames for IB nals (5602)
5563 - 2.6 changed lock ordering of 2 semaphores, caused deadlock (5654)
5564 - don't start multiple acceptors for the same port (5277)
5565 - fix incorrect LASSERT in mds_getattr_name (5635)
5566 - export a proc file for general "ping" checking (5628)
5567 - fix "lfs check" to not block when the MDS is down (5628)
5569 - service request history (4965)
5570 - put {ll,lov,osc}_async_page structs in a single slab (4699)
5571 - create an "evict_client" /proc entry on OSTs, like the MDS has
5572 - fix mount usage message, return errors per mount(8) (5168)
5573 - change grep [] to grep "[]" in tests so they work in more UMLs
5574 - fix ppc64/x86_64 spec to use %{_libdir} instead of /usr/lib (5389)
5575 - remove ancient LOV_MAGIC_V0 EA support (5047)
5576 - add "disk I/Os in flight" and "I/O req time" stats in obdfilter
5577 - align r/w RPCs to PTLRPC_MAX_BRW_SIZE boundary for performance (3451)
5578 - allow readahead allocations to fail when low on memory (5383)
5579 - mmap locking landed again, after considerable improvement (2828)
5580 - add get_hostaddr() to lustreDB.py for LDAP support (5459)
5582 2004-11-23 Cluster File Systems, Inc. <info@clusterfs.com>
5585 - send OST transaction number in read/write reply to free req (4966)
5586 - don't ASSERT in ptl_send_rpc() if we run out of memory (5119)
5587 - lock /proc/sys/portals/routes internal state, avoiding oops (4827)
5588 - the watchdog thread now runs as interruptible (5246)
5589 - flock/lockf fixes (but it's still disabled, pending 5135)
5590 - don't use EXT3 constants in llite code (5094)
5591 - memory shortage at startup could cause assertion (5176)
5593 - reorganization of lov code
5594 - single portals codebase
5596 - add extents/mballoc support (5025)
5597 - direct I/O reads in the obdfilter (4048)
5598 - kernel patches from LNXI for 2.6 (bluesmoke, perfctr, mtd, kexec)
5600 tbd Cluster File Systems, Inc. <info@clusterfs.com>
5603 - send OST transaction number in read/write reply to free req (4966)
5604 - don't ASSERT in ptl_send_rpc() if we run out of memory (5119)
5605 - lock /proc/sys/portals/routes internal state, avoiding oops (4827)
5606 - the watchdog thread now runs as interruptible (5246)
5607 - handle missing objects in filter_preprw_read properly (5265)
5608 - unsafe list practices (rarely) led to infinite eviction loop (4908)
5609 - drop import inflight refcount on signal_completed_replay error (5255)
5610 - unlock page after async write error during send (3677)
5611 - return original error code on reconstructed replies (3761)
5612 - no transno return for symlink open, don't save no-trasno open (3440)
5614 - add pid to ldlm debugging output (4922)
5615 - bump the watchdog timeouts -- we can't handle 30sec yet
5616 - extra debugging for orphan dentry/inode bug (5259)
5618 2004-11-16 Cluster File Systems, Inc. <info@clusterfs.com>
5621 - fix TCP_NODELAY bug, which caused extreme perf regression (5134)
5622 - allocate qswnal tx descriptors singly to avoid fragmentation (4504)
5623 - don't LBUG on obdo_alloc() failure, use OBD_SLAB_ALLOC() (4800)
5624 - fix NULL dereference in /proc/sys/portals/routes (4827)
5625 - allow failed mdc_close() operations to be interrupted (4561)
5626 - stop precreate on OST before MDS would time out on it (4778)
5627 - don't send partial-page writes before EOF from client (4410)
5628 - discard client grant for sub-page writes on large-page clients (4520)
5629 - don't free dentries not owned by NFS code, check generation (4806)
5630 - fix lsm leak if mds_create_objects() fails (4801)
5631 - limit debug_daemon file size, always print CERROR messages (4789)
5632 - use transno after validating reply (3892)
5633 - process timed out requests if import state changes (3754)
5634 - update mtime on OST during writes, return in glimpse (4829)
5635 - add mkfsoptions to LDAP (4679)
5636 - use ->max_readahead method instead of zapping global ra (5039)
5637 - don't interrupt __l_wait_event() during strace
5639 - add software watchdogs to catch hung threads quickly (4941)
5640 - make lustrefs init script start after nfs is mounted
5641 - fix CWARN/ERROR duplication (4930)
5642 - return async write errors to application if possible (2248)
5643 - add /proc/sys/portal/memused (bytes allocated by PORTALS_ALLOC)
5644 - print NAL number in %x format (4645)
5645 - update barely-supported suse-2.4.21-171 series (4842)
5646 - support for sles 9 %post scripts
5647 - support for building 2.6 kernel-source packages
5648 - support for sles km_* packages
5650 2004-10-07 Cluster File Systems, Inc. <info@clusterfs.com>
5653 - ignore -ENOENT errors in osc_destroy (3639)
5654 - notify osc create thread that OSC is being cleaned up (4600)
5655 - add nettype argument for llmount in #5d in conf-sanity.sh (3936)
5656 - reconstruct ost_handle() like mds_handle() (4657)
5657 - create a new thread to do import eviction to avoid deadlock (3969)
5658 - let lconf resolve symlinked-to devices (4629)
5659 - don't unlink "objects" from directory with default EA (4554)
5660 - hold socknal file ref over connect in case target is down (4394)
5661 - allow more than 32000 subdirectories in a single directory (3244)
5662 - fix blocks count for O_DIRECT writes (3751)
5663 - OST returns ENOSPC from object create when no space left (4539)
5664 - don't send truncate RPC if file size isn't changing (4410)
5665 - limit OSC precreate to 1/2 of value OST considers bogus (4778)
5666 - bind to privileged port in socknal and tcpnal (3689)
5668 - rate limit CERROR/CWARN console message to avoid overload (4519)
5669 - GETFILEINFO dir ioctl returns LOV EA + MDS stat in 1 call (3327)
5670 - basic mmap support (3918)
5671 - kernel patch series update from b1_4 (4711)
5673 2004-09-16 Cluster File Systems, Inc. <info@clusterfs.com>
5676 - avoid crash during MDS cleanup with OST shut down (2775)
5677 - fix loi_list_lock/oig_lock inversion on interrupted IO (4136)
5678 - don't use bad inodes on the MDS (3744)
5679 - dynamic object preallocation to improve recovery speed (4236)
5680 - don't hold spinlock over lock dumping or change debug flags (4401)
5681 - don't zero obd_dev when it is force cleaned (3651)
5682 - print grants to console if they go negative (4431)
5683 - "lctl deactivate" will stop automatic recovery attempts (3406)
5684 - look for existing locks in ldlm_handle_enqueue() (3764)
5685 - don't resolve lock handle twice in recovery avoiding race (4401)
5686 - revalidate should check working dir is a directory (4134)
5688 - don't always mark "slow" obdfilter messages as errors (4418)
5690 2004-08-24 Cluster File Systems, Inc. <info@clusterfs.com>
5693 - don't close LustreDB during write_conf until it is done (3860)
5694 - fix typo in lconf for_each_profile (3821)
5695 - allow dumping logs from multiple threads at one time (3820)
5696 - don't allow multiple threads in OSC recovery (3812)
5697 - fix debug_size parameters (3864)
5698 - fix mds_postrecov to initialize import for llog ctxt (3121)
5699 - replace config semaphore with spinlock (3306)
5700 - be sure to send a reply for a CANCEL rpc with bad export (3863)
5701 - don't allow enqueue to complete on a destroyed export (3822)
5702 - down write_lock before checking llog header bitmap (3825)
5703 - recover from lock replay timeout (3764)
5704 - up llog sem before sending rpc (3652)
5705 - reduce ns lock hold times when setting kms (3267)
5706 - change a dlm LBUG to LASSERTF, to maybe learn something (4228)
5707 - fix NULL deref and obd_dev leak on setup error (3312)
5708 - replace some LBUG about llog ops with error handling (3841)
5709 - don't match INVALID dentries from d_lookup and spin (3784)
5710 - hold dcache_lock while marking dentries INVALID and hashing (4255)
5711 - fix invalid assertion in ptlrpc_set_wait (3880)
5713 - add libwrap support for the TCP acceptor (3996)
5714 - add /proc/sys/portals/routes for non-root route listing (3994)
5715 - allow setting MDS UUID in .xml (2580)
5716 - print the stack of a process that LBUGs (4228)
5718 2004-07-14 Cluster File Systems, Inc. <info@clusterfs.com>
5721 - don't cleanup request in ll_file_open() on failed MDS open (3430)
5722 - make sure to unset replay flag from failed open requests (3440)
5723 - if default stripe count is 0, use OST count for inode size (3636)
5724 - update parent mtime/ctime on client for create/unlink (2611)
5725 - drop dentry ref in ext3_add_link from open_connect_dentry (3266)
5726 - free recovery state on server during a forced cleanup (3571)
5727 - unregister_reply for resent reqs (3063)
5728 - loop back devices mounting and status check on 2.6 (3563)
5729 - fix resource-creation race that can provoke i_size == 0 (3513)
5730 - don't try to use bad inodes returned from MDS/OST fs lookup (3688)
5731 - more debugging for page-accounting assertion (3746)
5732 - return -ENOENT instead of asserting if ost getattr+unlink race (3558)
5733 - avoid deadlock after precreation failure (3758)
5734 - fix race and lock order deadlock in orphan handling (3450, 3750)
5735 - add validity checks when grabbing inodes from l_ast_data (3599)
5737 - add /proc/.../recovery_status to obdfilter (3428)
5738 - lightweight CDEBUG infrastructure, debug daemon (3668)
5739 - change default OSC RPC parameters to be better on small clusters
5740 - turn off OST read cache for files smaller than 32MB
5741 - install man pages and include them in rpms (3100)
5742 - add new init script for (un)mounting lustre filesystems (2593)
5743 - run chkconfig in %post for init scripts (3701)
5744 - drop scimac NAL (unmaintained)
5746 2004-06-17 Cluster File Systems, Inc. <info@clusterfs.com>
5749 - clean kiobufs before and after use (3485)
5750 - strip trailing '/'s before comparing paths with /proc/mounts (3486)
5751 - remove assertions to work around "in-flight rpcs" recovery bug (3063)
5752 - change init script to fail more clearly if not run as root (1528)
5753 - allow clients to reconnect during replay (1742)
5754 - fix ns_lock/i_sem lock ordering deadlock for kms update (3477)
5755 - don't do DNS lookups on NIDs too small for IP addresses (3442)
5756 - re-awaken ptlrpcd if new requests arrive during check_set (3554)
5757 - fix cond_resched (3554)
5758 - only evict unfinished clients after recovery (3515)
5759 - allow bulk resend, prevent data loss (3570)
5760 - dynamic ptlrpc request buffer allocation (2102)
5761 - don't allow unlinking open directory if it isn't empty (2904)
5762 - set MDS/OST threads to umask 0 to not clobber client modes (3359)
5763 - remove extraneous obd dereference causing LASSERT failure (3334)
5764 - don't use get_cycles() when creating temp. files on the mds (3156)
5765 - hold i_sem when setting i_size in ll_extent_lock() (3564)
5766 - handle EEXIST for set-stripe, set proper directory name (3336)
5768 - servers can dump a log evicting a client - lustre.dump_on_timeout=1
5769 - fix ksocknal_fmb_callback() error messages (2918)
5771 2004-05-27 Cluster File Systems, Inc. <info@clusterfs.com>
5774 - don't copy lvb into (possibly NULL) reply on error (2983)
5775 - don't deref dentry after dput, don't free lvb on error (2922)
5776 - use the kms to determine writeback rpc length (2947)
5777 - increment oti_logcookies when osc is inactive (2948)
5778 - update client's i_blocks count via lvb messages (2543)
5779 - handle intent open/close of special files properly (1557)
5780 - mount MDS with errors=remount-ro, like obdfilter (2009)
5781 - initialize lock handle to avoid ASSERT on error cleanup (3057)
5782 - don't use cancelling-locks' kms values (2947)
5783 - use highest lock extent for kms, not last one (2925)
5784 - don't dereference ERR_PTR() dentry in error handling path (3107)
5785 - fix thread race in portals_debug_dumplog() (3122)
5786 - create lprocfs device entries at setup instead of at attach (1519)
5787 - common AST error handler, don't evict client on completion race (3145)
5788 - zero nameidata in detach_mnt in 2.6 (3118)
5789 - verify d_inode after revalidate_special is valid in 2.6 (3116)
5790 - use lustre_put_super() to handle zconf unmounts in 2.6 (3064)
5791 - initialize RPC timeout timer earlier for 2.6 (3219)
5792 - don't dereference NULL reply buffer if mdc_close was never sent (2410)
5793 - print nal/nid for unknown nid (3258)
5794 - additional checks for oscc recovery before doing precreate (3284)
5795 - fix ll_extent_lock() error return code for 64-bit systems (3043)
5796 - don't crash in mdc_close for bad permissions on open (3285)
5797 - zero i_rdev for non-device files (3147)
5798 - clear page->private before handing to FS, better assertion (3119)
5799 - tune the read pipeline (3236)
5800 - fix incorrect decref of invalidated dentry (2350)
5801 - provide read-ahead stats and refine rpc in flight stats (3328)
5802 - don't hold journal transaction open across create RPC (3313)
5803 - update atime on MDS at close time (3265)
5804 - close LDAP connection when recovering to avoid server load (3315)
5805 - update iopen-2.6 patch with fixes from 2399,2517,2904 (3301)
5806 - don't leak open file on MDS after open resend (3325)
5807 - serialize filter_precreate and filter_destroy_precreated (3329)
5808 - loop device shouldn't call sync_dev() for nul device (3092)
5809 - clear page cache after eviction (2766)
5810 - resynchronize MDS->OST in background (2824)
5811 - refuse to mount the same filesystem twice on same mountpoint (3394)
5812 - allow llmount to create routes for mounting behind routers (3320)
5813 - push lock cancellation to blocking thread for glimpse ASTs (3409)
5814 - don't call osc_set_data_with_check() for TEST_LOCK matches (3159)
5815 - fix rare problem with rename on htree directories (3417)
5817 - allow default OST striping configuration per directory (1414)
5818 - fix compilation for qswnal for 2.6 kernels (3125)
5819 - increase maximum number of MDS request buffers for large systems
5820 - change liblustreapi to be useful for external progs like lfsck (3098)
5821 - increase local configuration timeout for slow disks (3353)
5822 - allow configuring ldlm AST timeout - lustre.ldlm_timeout=<seconds>
5824 2004-03-22 Cluster File Systems, Inc. <info@clusterfs.com>
5827 - fixes for glimpse AST timeouts / incorrectly 0-sized files (2818)
5828 - don't overwrite extent policy data in reply if lock was blocked (2901)
5829 - drop filter export grants atomically with removal from device (2663)
5830 - del obd_self_export from work_list in class_disconnect_exports (2908)
5831 - don't LBUG if MDS recovery times out during orphan cleanup (2530)
5832 - swab reply message in mdc_close, other PPC fixes (2464)
5833 - fix destroying of named logs (2325)
5834 - overwrite old logs when running lconf --write_conf (2264)
5835 - bump LLOG_CHUNKSIZE to 8k to allow for larger clusters (2306)
5836 - fix race in target_handle_connect (2898)
5837 - mds_reint_create() should take same inode create lock (2926)
5838 - correct journal credits calculated for CANCEL_UNLINK_LOG (2931)
5839 - don't close files for self_export to avoid uninitialized obd (2936)
5840 - allow MDS with the same name as client node (2939)
5841 - hold dentry reference for closed log files for unlink (2325)
5842 - reserve space for all logs during transactions (2059)
5843 - don't evict page beyond end of stripe extent (2925)
5844 - don't oops on a deleted current working directory (2399)
5845 - handle hard links to targets without a parent properly (2517)
5846 - don't dereference NULL lock when racing during eviction (2867)
5847 - don't grow lock extents when lots of conflicting locks (2919)
5849 2004-03-04 Cluster File Systems, Inc. <info@clusterfs.com>
5852 - account for cache space usage on clients to avoid data loss (974)
5853 - lfsck support in lustre kernel code (2349)
5854 - reduce journal credits needed for BRW writes (2370)
5855 - orphan handling to avoid losing space on client/server crashes
5856 - ptlrpcd can be blocked, stopping ALL progress (2477)
5857 - use lock value blocks to assist in proper KMS, faster stat (1021)
5858 - takes i_sem instead of DLM locks internally on obdfilter (2720)
5859 - recovery for initial connections (2355)
5860 - fixes for mds_cleanup_orphans (1934)
5861 - abort_recovery crashes MDS in b_eq (mds_unlink_orphan) (2584)
5862 - block all file creations until orphan recovery completes (1901)
5863 - client remove rq_connection from request struct (2423)
5864 - conf-sanity test_5, proper cleanup in umount log not availale (2640)
5865 - recovery timer race (2670)
5866 - mdc_close recovey bug (2532)
5867 - ptlrpc cleanup bug (2710)
5868 - mds timeout on local locks (2588)
5869 - namespace lock held during RPCs (2431)
5870 - handle interrupted sync write properly (2503)
5871 - don't try to handle a message that hasn't been replied to (2699)
5872 - client assert failure during cleanup after abort recovery (2701)
5873 - leak mdc device after failed mount (2712)
5874 - ptlrpc_check_set allows timedout requests to complete (2714)
5875 - wait for inflight reqs when ptlrpcd finishes (2710)
5876 - make sure unregistered services are removed from the srv_list
5877 - reset bulk XID's when resending them (caught by 1138 test)
5878 - unregister_bulk after timeout
5879 - fix lconf error (2694)
5880 - handle write after unfinished setstripe, stripe-only getstripe (2388)
5881 - readahead locks pages, leaves pending causing memory pressure (2673)
5882 - increase OST request buffers to 4096 on large machines (2729)
5883 - fix up permission of existing directories in simple_mkdir (2661)
5884 - init deleted item, add assertions ptlrpc_abort_inflight() (2725)
5885 - don't assign transno to errored transactions (2742)
5886 - don't delete objects on OST if given a bogus objid from MDS (2751)
5887 - handle large client PAGE_SIZE readdir on small PAGE_SIZE MDS (2777)
5888 - if rq_no_resend, then timeout request after recovery (2432)
5889 - fix MDS llog_logid record size, 64-bit array alignment (2733)
5890 - don't call usermode_helper from ptlrpcd, DEFAULT upcall (2773)
5891 - put magic in mount.lustre data, check for bad/NULL mount data (2529)
5892 - MDS recovery shouldn't delete objects that it has given out (2730)
5893 - if enqueue arrives after completion, don't clobber LVB (2819)
5894 - don't unlock pages twice when trigger_group_io returns error (2814)
5895 - don't deref NULL rq_repmsg if ldlm_handle_enqueue failed (2822)
5896 - don't write pages to disk if there was an error (1450)
5897 - don't ping imports that have recovery disabled (2676)
5898 - take buffered bytes into account when balancing socknal conn (2817)
5899 - hold a DLM lock over readdir always, use truncate_inode_pages (2706)
5900 - reconnect unlink llog connection after MDS reconnects to OST (2816)
5901 - remove little-endian swabbing of llog records (1987)
5902 - set/limit i_blksize to LL_MAX_BLKSIZE on client (2884)
5903 - retry reposting request buffers if they fail (1191)
5904 - grow extent at grant time to avoid granting a revoked lock (2809)
5905 - lock revoke doesn't evict page if covered by a second lock (2765)
5906 - disable VM readahead to avoid reading outside lock extents (2805)
5908 - return LL_SUPER_MAGIC from statfs for the filesystem type (1972)
5909 - updated kernel patches for hp-2.4.20 kernel (2681)
5911 2004-02-07 Cluster File Systems, Inc. <info@clusterfs.com>
5914 - fix truncated write corruption (2366)
5915 - fix for failed assertion in iopen_connect_dentry (1792,2517)
5917 - don't flag the ptlrpcd thread with PF_MEMALLOC (2636)
5918 - ensure len(uuid) < 37 in lmc (1171)
5919 - fix ia64 OOPS in llog_test (2255)
5920 - zero end of page at obdfilter for partial page writes (2648)
5921 - don't leave stale dentries around after renames (bug 2428)
5922 - fix timeouts when evicting a client with a single lock held (2642)
5923 - set deadline for the initial HELLO message to drain (2634)
5924 - print out dotted-quad IP addresses in the socknal (2302)
5926 - additional debugging for MDS client eviction problem (2443)
5927 - fix mkfsoptions support for osts (2603, 2604)
5929 2004-01-27 Cluster File Systems, Inc. <info@clusterfs.com>
5932 - add series for the vanilla 2.6.0 kernel
5933 - add series for the vanilla 2.4.24 kernel
5934 - add series for a cray x86/64 UL kernel drop
5935 - fix xattr patches for the vanilla 2.4.19 series
5937 - generate true UUIDs in lmc (1171)
5938 - have portals stack dumping break in UML (2466)
5939 - avoid bad dchild deref; avoid inum lock w/o creation (2362)
5940 - allocate with _NOFS in ldlm to avoid deadlock (1933)
5941 - wake callback waiting threads on client eviction (2460)
5942 - Add --ptldebug and --subsystem to lmc (1719)
5943 - update assertion to allow safe interrupt allocation
5944 - set rq_no_resend for cancel requests (2432)
5945 - recalculate ptlrpcd timeout after resend (2494)
5946 - call vfs_rmdir when removing pending directories (2368)
5947 - fix renaming a file to itself (2429)
5948 - lmc creates a default one-stripe lov (2454)
5949 - expand procfs space to handle large clusters (2326)
5950 - increase UML stack to avoid overflow
5951 - update lconf's list of debug and subsystem masks
5952 - fix lfs find --obd (2510)
5953 - /proc tunable for disabling filter read caching (2591)
5954 - stop rpm packages from altering slapd.conf (2301)
5955 - disable nagle in the socknal under 0conf (2578)
5956 - choose mds inode size based on stripe count (2572)
5957 - fix kernel-source rpm problems (2516)
5959 - add --disable-doc to avoid pdf generation (2421)
5960 - update documentation, tests, type-os, comments
5961 - avoid format warnings on ia64
5962 - remove the TOE NAL
5963 - tiny code cleanups by removing unused fields
5965 2004-01-07 Cluster File Systems, Inc. <info@clusterfs.com>
5968 - fix obvious semaphore misuse in as-yet-unused setattr path (2348)
5969 - remove the most blatant lies from BUILDING file (2371)
5970 - change default debug level to reasonable production setting
5971 - reduce client side cache size to reduce cache flush time
5972 - reduce max RPCs in flight to avoid unnecessary file fragmentation
5973 - make TCP zerocopy and pinger support enabled by default (2476)
5974 - sync writes completed after process exits caused crashes (2319)
5975 - maintain correct mount count on the MDS (2356)
5976 - backout 1557, because 2316 wasn't really fixed
5977 - better file I/O statistics gathering in /proc
5978 - don't take unnecessary, deadlock-inducing bug in readpage (2383)
5979 - another kernel patch to fix zero-copy TCP function export
5980 - don't take duplicate lock when processing re-sent getattr (2420)
5981 - lctl uses obd_self_export instead of creating new conn (2353)
5982 - MDS/OST recovery case which requires object creation asserted (2425)
5983 - move lfs from /usr/sbin to /usr/bin in packages
5984 - fix race between mds_client_add and mds_client_free (2417)
5985 - use kmalloc instead of slabs in portals (2430)
5986 - don't create duplicate records when a failover MDS is present (2442)
5987 - remove unnecessary mount age check (2332)
5988 - don't remove directory inodes from locks prematurely (2451)
5989 - don't break if MDS service name is the same as hostname (2103)
5990 - fix races in client write RPC generation when cache full (2482)
5992 2003-12-13 Cluster File Systems, Inc. <info@clusterfs.com>
5995 - remove now-unused request->rq_obd (278)
5996 - if an allocation fails, print out how much memory we've used (1933)
5997 - use PORTAL_SLAB_ALLOC for structures, to get GFP_MEMALLOC (1933)
5998 - add the "configurable stack size" patch to most series files (1256)
5999 - ability to write large log records, for 100+ OST configs (2306)
6000 - fix NULL deref when filter_prep fails (2314)
6001 - fix operator precedence error in filter_sync
6002 - dynamic allocation of socknal TX descriptors (2315)
6003 - fix a missed case in the GFP_MEMALLOC patch, can cause deadlock (2310)
6004 - fix gcc 2.96 compilation problem in xattr kernel patch (2294)
6005 - ensure that CWARN messages in Portals always get to the syslog
6006 - __init/__exit are not for prototype decls (ldlm_init/exit)
6007 - x86-64 compile warning fixes
6008 - fix gateway LMC keyword conflict (2318)
6009 - fix MDS lock inversions in getattr/reint paths (1844)
6010 - fix a rare lock re-ordering bug, which caused deadlock (2322)
6011 - fix i_sem/journal inversion in fsfilt_ext3_write_record (2306)
6012 - DLM race condition prevented some lock evictions (2328)
6013 - ENOMEM detection and retry on socknal sends (2230)
6014 - use GFP_NOFS throughout Lustre, to combat ENOMEM (2230)
6015 - move osc_rpcd into ptlrpc, for use in MDC and others (2329)
6016 - protect MDS inode fsdata with stronger locking; fixes assertion (2313)
6017 - better error messages when a client is rejected during recovery (1505)
6018 - avoid cancelling locks which were never granted, after failure (2330)
6019 - fix i_sem/journal inversion in mds_client_add (2333)
6020 - fix truncate/getattr lock cycle deadlock (2334)
6021 - use rpcd to send close; allows resend after timeout, avoid leak (1897)
6022 - fix two rare exit paths which could leak an l_lock() ref (2321)
6023 - fencepost error in MDS/OST orphan recovery (2226)
6024 - make log record alignment 8 bytes (1988)
6025 - lstripe now fails when requested offset > ost_count (2237)
6026 - ensure that all kernel series have a complete list.h (1607)
6027 - fix crashes in special-file operations (2316)
6028 - lctl create/brw OID mismatch, caused by obsolete filter loop (2339)
6030 - allow configurable automake binary, for testing new versions
6031 - small update to the lfs documentation
6033 2003-12-03 Cluster File Systems, Inc. <info@clusterfs.com>
6035 * fix negative export reference count in fsfilt_sync (2312)
6037 2003-12-01 Cluster File Systems, Inc. <info@clusterfs.com>
6038 * release candidate 0.9.1
6040 - orphans are moved into the PENDING directory for possible recovery
6041 - replayed opens now open by fid for orphan/rename safety (1042)
6042 - last close of an orphan inode generates a transno (683)
6043 - chdir() and mount() now pin the directory entry (1020)
6044 - avoid CERROR in normal ll_setattr_raw() error case (1500)
6045 - discard very old requests without processing them (1502)
6046 - remove some common, well-understood CERRORs (1505)
6047 - require O_DIRECT I/O to be page-sized to workaround IA64 crash (1609)
6048 - clear "grant" flags in OST replies until OST grant code lands (1644)
6049 - fix read performance by not clobbering i_blksize on client (1598)
6050 - fix __ldlm_handle2lock oops by not dereferencing lock after PUT (1625)
6051 - make LRU size a /proc tunable, clears locks when reduced (707)
6052 - fix some lprocfs rot that prevented ptlbd from loading (1732)
6053 - server locks take references on exports now (1558)
6054 - build fixes for 2.4.20-rh trees (1663)
6055 - return an error from lov_create if all OSCs are inactive (1751)
6056 - fix import levels when a reconnect happens without a timeout (1597)
6057 - exit early from mds_open if we get a lookup error (1749)
6058 - partial page read at EOF wouldn't wait for disk before sending (1642)
6059 - avoid NULL deref in obdfilter when reading page past EOF (1592)
6060 - avoid LASSERT in ll_intent_lock if server failed very early (1090)
6061 - fix LBUG in ll_it_open_error with rc = -2 (1861)
6062 - write/truncate lock inversion (1639)
6063 - Don't auto-load obdclass, portals modules during cleanup (1495)
6064 - fix timestamps from jumping to "now" (1763)
6065 - extra journal assertions (1648)
6066 - add an extra multiunlink test (1771)
6067 - fix read_record/write_record API (1776)
6068 - fix leak of offset_extent, possible incorrect i_size later (1772)
6069 - fix lasserts in mis-matched transnos during open-unlink testing (1541)
6070 - Debugging for the kqswnal_get_idle_tx problems (1820)
6071 - Allow recovery to be attempted multiple times (1536)
6072 - Write out MDS last_rcvd file after it is first created (1600)
6073 - Fix tx_descriptor leak in failed transmit situations (1827)
6074 - ext3 journaling fixes for assertion failure after IO error (1871)
6075 - class_export_put() on freed export after completion AST error (1896)
6076 - Fix revalidate looping in VFS (1322)
6077 - Don't access a freed export during MDS_REINT timeout (1521)
6078 - Add open-unlink recovery support on the MDS (1673,1764)
6079 - Return an error if no MDS data was read from last_rcvd (1946)
6080 - Fix for lookup "." or ".." crash on error (1932,1931,1935)
6081 - Don't setup a disk device that doesn't match exported UUID (317)
6082 - Reduce bulk RPC timeout to avoid cascading client/OST failures (1845)
6083 - avoid committing NULL handle in force close
6084 - local.sh is now a one-stripe LOV configuration
6085 - POSIX utime.4 -EPERM on FIFO not owned by user (56)
6086 - fix ext3 htree duplicate directory entry corruption (1516)
6087 - POSIX creat.13, fstat.1, open.18, stat.3 new file atime/mtime (2020)
6088 - update to new LOV EA format (2097)
6089 - interoperability for different PAGE_SIZE/wordsize (686,1821,1343,2042)
6091 2003-06-15 Phil Schwan <phil@clusterfs.com>
6094 - imports and exports cleanup too early, need refcounts (349, 879, 1045)
6095 - per-import/export recovery handling (958, 931, 959)
6096 - multiple last-rcvd slots, for serving multiple FSes (949)
6097 - connections are again shared between multiple imp/exports (963, 964)
6098 - "umount -f" would hang if any requests needed to be sent (393, 978)
6099 - avoid pinning large req buffer by copying for queued messages (989)
6100 - add "uuid" to "lctl device" command to help upcalls (991)
6101 - "open" RPCs with transnos would confuse recovery counters (1037)
6102 - do proper endian conversion of all wire messages (288, 340, 891)
6103 - remove OST bulk get LBUGs, fix ost_brw_write cleanup (1126)
6104 - call waiting locks callback from LDLM recovery thread (1127, 1151)
6105 - fix ptlrpc_connection leak in target_handle_connect (1174)
6106 - fix import refcounting bug in OST and MDS cleanup (1134)
6107 - if an invalid-at-open-time OSC returned before close(), LBUG (1150)
6108 - fix very unlikely obd_types race condition (501)
6109 - remove osc_open hack for echo_client (1187)
6110 - we leaked exports/dlmimps for forcibly disconnected clients (1143)
6111 - a failure in read_inode2 leads to deadlock (1139)
6112 - cancel ack-locks as soon as transaction is committed (1072)
6113 - fix major leaks and crashes in the bulk I/O path (937, 1057)
6114 - make sure to commitrw after any preprw to avoid deadlock (1162)
6115 - failing to execute a file in a lustre FS would lock inode (1203)
6116 - small DEBUG_REQ fix to avoid dereferencing a NULL (1227)
6117 - don't ASSERT while cleaning up an incompletely-setup obd (1248)
6118 - obd_uuid2tgt would walk off the end of the list (1255)
6119 - on IA64 the osc would give portals incorrect bulk size (1258)
6120 - fix debug daemon ioctl interface; allows daemon on ia64 (1274)
6121 - fix lock inversion caused by new llite matching code (1282)
6122 - limit the number of dirty pages on a client to 10MB (1286)
6123 - timed out locks were not being corrected cancelled (1289)
6124 - fix O_DIRECT above 4GB on IA-32 (1292)
6125 * major user-visible changes
6126 - fail out/fail over policy now controlled by the upcall (993)
6128 - add OBD_PING to check server availability and failure (954)
6129 - lustre messages are now sent in sending host order (288, 340, 891)
6130 - add eadatalen to MDS getattr reply (340)
6131 - OST read replies may contain second buffer, with per-page status (593)
6133 2003-03-11 Phil Schwan <phil@clusterfs.com>
6136 - LDLM_DEBUG macro fix, for gcc 3.2 (850)
6137 - failed open()s could cause deadlock; fixed (867, 869)
6138 - stop cancelling OST locks when files are closed (481)
6139 - overlapping XID spaces caused network corruption (851, 853)
6140 - fix unsafe fsfilt counter arithmetic; change to atomic_t
6141 - setattr_raw added, to do single-RPC, server-side setattrs
6142 - lmc/lconf syntax change for OST UUIDs
6143 - fix crashy race condition between ptlrpc_free_req and osc_close
6144 - don't use request in mdc_enqueue if we hit a timeout (889)
6145 - don't set the inode i_size for regular files from the MDS (896)
6146 - handle out of order completion AST (842)
6147 - don't LBUG if a lock request times out after receiving AST (913)
6148 - avoid d_rehash race in ll_find_alias by rehashing inside dcache_lock
6149 - if a bad lock AST arrives, send an error instead of dropping entirely
6150 - return 0 from revalidate2 if ll_intent_lock returns -EINTR (912)
6151 - fix leak in bulk IO when only partially completed (899, 900, 926)
6152 - fix O_DIRECT for ia64 (55)
6153 - (almost) eliminate Lustre-kernel-thread effects on load average (722)
6154 - C-z after timeout could hang a process forever; fixed (977)
6156 - client-side I/O cache (678, 924, 929, 941, 970)
6158 - READPAGE and SETATTRs which don't take server-side locks get
6161 2003-02-11 Phil Schwan <phil@clusterfs.com>
6164 - Fix ldlm_lock_match on the MDS to avoid matching remote locks (592)
6165 - Fix fsfilt_extN_readpage() to read a full page of directory
6166 entries, or fake the remainder if PAGE_SIZE != blocksize (500)
6167 - Avoid extra mdc_getattr() in ll_intent_lock when possible (534, 604)
6168 - Fix imbalanced LOV object allocation and out-of-bound access (469)
6169 - Most intent operations were removed, in favour of a new RPC mode
6170 that does a single RPC to the server and bypasses most of the VFS
6171 - All LDLM resource ID arrays were removed in favour of ldlm_res_id
6172 - Aggressively cancel local locks on DLM servers
6173 - mds_reint_unlink sends EA to the client if it's the last nlink.
6174 client uses that EA to unlink OST objects.
6175 - mds_reint_{rename,unlink,link} were rewritten to take ordered locks
6176 - recursive symlinks were fixed (439)
6177 - fixed NULL deref in DEBUG_REQ
6178 - filter_update_lastobjid no longer calls sync, which annoyed extN
6179 - fixed multi-client small-writes to a single file problem (445)
6180 - fixed mtime updates during file writes (607)
6181 - fixed vector writes on obdfilter causing problems when ENOSPC (670)
6182 - fixed bug in obd_brw_read/write() (under guise of testing 367)
6183 - fixed Linux OST size reporting problem (444, 656)
6184 - OST now updates object mtime with writes or setattr (607, 619)
6185 - client verifies file size before zeroing page past EOF (445)
6186 - OST now writes last allocated objid to disk with allocation (108)
6187 - LOV on echo now works (409)
6189 - mds_reint_unlink sends a new buffer, with the EA included. this
6190 buffer is only valid if body->valid & OBD_MD_FLEASIZE, which is only
6191 set if a regular file was being unlinked, and it was the last link
6192 - use PtlGet from the target for bulk writes (315)
6193 - OST now updates object mtime with writes or setattr (607, 619)
6194 - LDLM now has a grant-time callback to revalidate locked items, if
6196 - Many MDS operations were reorganized to combat race conditions
6198 - Merge b_intel branch (updated lprocfs code) - now at /proc/fs/lustre
6199 - configure check to avoid gcc version 2.96 20000731-2.96-98) (606)
6201 2003-01-06 Andreas Dilger <adilger@clusterfs.com>
6204 - Fully reactivate OST imports after reconnection (512, others)
6205 - Make sure client sees our -ENOTCONN from mds_handle (513 - partial)
6206 - More graceful error handling for truncating on dead OST (515)
6207 - Don't error out unless we're actually accessing dead stripes (474)
6208 - Fix garbage sizes when stripes are missing (410)
6209 - LRU counters were broken, causing constant lock purge (433, 432)
6210 - garbage on read from stripes with failed OSTs (441)
6211 - mark OSCs as active before reconnecting during recovery (438)
6212 - lov_enqueue and lov_cancel need to handle inactive OSTs (403)
6213 - lfind did not preserve OST order in output (443)
6214 - symlinks cause hung clients, incorrect data (439)
6215 - stop dereferencing request after dropping refcount (457)
6216 - don't LASSERT(spin_is_locked) on non-SMP (455)
6217 - fixes for many rename() bugs
6218 - fstat didn't correctly synchronize attributes (399)
6219 - server must handle lock cancellation during blocking AST prep (487)
6220 - bulk descriptors were free()d too soon (511)
6221 - fix paths in lconf, which would load incorrect modules (451, 507)
6222 - fix confusing lconf 'host not found' error message (386)
6223 - fix lock order deadlock on OST (O/R i_sem before journal ops, 478)
6224 - fix race condition in mdc_blocking_ast() for inode access (526)
6225 - fix lov_unpackmd() unpacking wrong number of stripes (537)
6226 - fix lov_set_osc_active() marking wrong OSC inactive (440)
6227 - fix bad lstripe lov_unpackmd() assertion (fix layering too) (527)
6228 - fix multiple writes of stripe MD to MDS (358, maybe 519)
6229 - fix lstripe in several ways (kernel side) (527)
6230 - fix request leak in ldlm_cli_enqueue (262)
6231 - incorrect OSC was marked inactive after OST failure
6232 - call mds_fs_cleanup before unmounting filesystem (524)
6233 - fix races between taking ns_lock and ldlm_lock_change_resource
6234 - fix races updating LOV export open file list
6235 - fix lov_enqueue error path, avoid decref-ing bad lock handle
6236 - fix recovery NULL deref in ldlm_cli_cancel_unused
6237 - fix some DLM races by using new hash table for lock handles (419)
6238 - permit the client to specify desired inodes, at replay
6239 - duplicate requests when we queue them for replay reintegration
6240 - fix last_rcvd offset calculation
6241 - sync after each recovered transaction, so we always make progress
6242 - never, not always, ERESTART requests without transnos
6243 - store the lov_desc in the MDS, so we don't depend on getlovinfo to
6245 - skip replay if the MDS says that the client is already connected
6246 - don't check for a recovery-enabled export to match lctl's UUID
6247 - don't INC_USE_COUNT for phantom exports
6248 - don't crash when cleaning up phantom exports (567)
6249 - don't double-finish or set replay data for errored mdc_open requests
6250 - abort requests when they time out, so we don't get old replies
6251 - send/receive replies for AST messages again
6252 - if the client says that it doesn't have the lock, cancel it on the
6254 - if we timeout during I/O, don't try to cancel an in-use lock; instead
6255 mark it as destroyed, it will all work out when decref is called
6256 - fix module use counts (22, 581)
6258 - ASTs now expect a reply (server cancels lock on error reply)
6260 2002-12-02 Andreas Dilger <adilger@clusterfs.com>
6263 - fix many simultaneous client startup (392)
6264 - fix dentry->d_it clobbering
6265 - credentials weren't being shipped for readdir/getattr operations
6266 - remove invalid assertions triggered during some concurrent MD
6268 - proper Lustre versions added (336, 389)
6269 - fix memory leak for create error case (398)
6270 - fix LOV locking bug that would get cli/srv out of sync
6271 - fix echo client over LOV (409)
6272 - fix dbench 2, extN refcount problem (170, 258, 356, 418)
6273 - fix double-O_EXCL intent crash (424)
6274 - avoid sending multiple lock CANCELs (352)
6276 - MDS can do multi-client recovery (modulo bugs in new code)
6278 - many updates, edits, cleanups
6280 2002-11-18 Phil Schwan <phil@clusterfs.com>
6283 - fix null d_it dereference (346)
6284 - fix full OST/dbench hang (333)
6285 - fix permission problem with file removal (286)
6286 - fix removal of OSCs from LOV when they fail
6287 - fix NULL deref during bulk timeout (214)
6288 - fix problems related to multiple filesystems on one MDS (241)
6289 - fixed serious subtle metadata locking bugs
6290 - free locks on clients when inodes are removed due to memory
6292 - fix inode pointer in lock data (285)
6293 - partial support for multiple MDS on a single host (241)
6294 - data locks weren't cancelled at clear_inode time (290, 311)
6295 - intent locks could lead to unbounded lock growth (205)
6296 - added a maximum lock count, an LRU list, and a flusher
6297 - fix multiple rename (365)
6298 - properly abstracted the echo client
6299 - OSC locked 1 byte too many; fixed
6300 - rewrote brw callback code:
6301 - fixed recovery bugs related to LOVs (306)
6302 - fixed too-many-pages-in-one-write crash (191)
6303 - fixed (again) crash in sync_io_timeout (214)
6304 - probably fixed callback-related race (385)
6306 - Add capability to MDS protocol
6307 - LDLM cancellations and callbacks on different portals
6309 2002-10-28 Andreas Dilger <adilger@clusterfs.com>
6312 - limit client IOV size to PTL_MD_MAX_IOV (611336, 191)
6313 - defer open object destruction to close time (601981, 138)
6314 - open/close OST file handle in obdo (OBD_MD_FLHANDLE) (601981, 138)
6315 - move LDLM_ENQUEUE/CONVERT back to MDS portal (625069)
6316 - abstract ll_lookup2, fix ll_revalidate2 to use abstraction (256)
6317 - don't call obd_setattr in ll_file_release for destroyed objects
6318 * protocol change to lustre_msg: move |version| and add |flags|
6319 * protocol change to osc_punch: "start" in "o_size", "end" in "o_blocks"
6320 * lock replay: for LDLM_FL_REPLAY trust client to do right thing
6321 * added replay of create, unlink, link and rename operations during
6322 MDS failover; recovery should be much more robust now
6323 * remove failed OSCs from LOVs (only lov_create uses this so far)
6324 * the lustre-HOWTO was brought (more) up to date (582544)
6326 2002-10-23 Phil Schwan <phil@clusterfs.com>
6329 - in-use dentries weren't being reused properly (617851)
6330 - prevent multiple LDLM setup (599178)
6331 - fix LOV size calculations for truncate (617853)
6332 - fix client handling of MDS intent errors (POSIX)
6333 - fix permission bug in lovstripe.c test (624321)
6334 - fix MDS thread deadlock - move LDLM handler to DLM portal (625069)
6335 - truncate past end of file could corrupt data
6336 - proper cleanup after timeouts, crashes, etc (592524, 550815)
6337 - a race in recovery could return ETIMEDOUT to apps (623947)
6338 - building outside the source directory was fixed
6339 * the lustre-HOWTO was brought (more) up to date (582544)
6340 * major progress was made on recovery functionality
6342 2002-10-10 Phil Schwan <phil@clusterfs.com>
6345 - recovery deadlock fix
6346 - rm -rf causes LBUG fix (617817)
6347 - file open by multiple tasks fix (618962)
6348 - directory permissions bugs (602707 and 620007)
6349 - journal_stop fixed with locking (611313)
6350 - O_APPEND failures resolved (618273, perhaps 614459)
6351 - lconf PATH fix (619770)
6352 - IA64 build fix (621450)
6353 - RPC buffer sizes scale with amount of memory
6355 2002-10-01 Phil Schwan <phil@clusterfs.com>
6358 - locks would be cancelled without throwing away data pages,
6359 resulting in inconsistent data (605627)
6360 - inode attributes were not always being refreshed (605627, 612449)
6361 - lconf now continues to cleanup after lctl reports an error
6362 - MDS now enforces user permissions (602707)
6363 - lprocfs cleanup fixed, but not yet enabled (614157)
6364 - fixed infinite server hang, should a client not respond to an AST
6365 - avoid going into recovery if user calls readlink() with a buffer
6366 that's too small (613941)
6367 - AST RPCs no longer require replies (614867) -- this may be changed
6368 - don't crash server if client sends an IOV that's too big (611336)
6369 - fixed lock conversion deadlock (611892)
6370 - fixed the following of symlinks (614622)
6371 * recovery: the server can remove locks from a client that dies, other
6372 clients can make progress
6373 * more extN patch fixes
6374 * compile-time configurable ptlrpc buffer allocations
6376 - collaborative read cache document
6377 - Lustre Lite Performance CDR document-in-progress
6379 2002-09-20 Andreas Dilger <adilger@clusterfs.com>
6382 - fix typo in patch-2.4.18
6384 2002-09-20 Andreas Dilger <adilger@clusterfs.com>
6387 - clear ptlrpc request each time in handle_incoming_request()
6388 - unlink of files now destroys the object on the OST
6390 2002-09-19 Peter Braam <braam@clusterfs.com>
6392 * add hard link support
6393 * change obdfile creation method
6394 * kernel patch changed
6396 2002-09-19 Peter Braam <braam@clusterfs.com>
6399 - stack overflow bug in extN fixed
6401 2002-09-18 Andreas Dilger <adilger@clusterfs.com>
6403 * documentation updates
6404 - add man pages for config tools
6405 - update tests/README to describe testing with new config tools
6406 - finish metadata API descriptions
6407 * bug fixes and cleanups
6408 - statfs workaround for 16TB limit
6409 - LOV stripe allocation improved, can stripe on subset of OSTs
6410 - LOV file size/IO offset was wrong for files > 4GB in size
6411 - object EA data was being dropped, caused files to be unreadable
6412 - memory overflow with non-LOV OST caused memory corruption
6413 - fixed regression tests to work with new config tools, obdfilter
6414 - fixed bug when directory size became larger than 1 block
6415 - fixed bug (for single client case) when PWD was deleted
6416 - invalidate local directory pages when doing intent-based ops
6417 - avoid LDLM oops when lock callback contained bad data
6419 2002-09-09 Andreas Dilger <adilger@clusterfs.com>
6421 * documentation updates
6422 * bug fixes and cleanups
6423 - configuration tools
6426 - 64-bit compile warnings
6427 - 64-bit internal statfs data
6429 * test_brw on persistent OST devices
6431 * lprocfs (disabled)
6433 2002-09-04 Andreas Dilger <adilger@clusterfs.com>
6435 * documentation updates
6436 * bug fixes and cleanups
6437 * configuration tools
6439 2002-08-30 Peter J. Braam <braam@clusterfs.com>
6442 * many small fixes to 0_5_4
6443 * io/network handling
6444 * thinkos in MDS operations
6446 2002-08-24 Peter J. Braam <braam@clusterfs.com>
6449 * crucial basic fixes to 0.5.3
6450 * IOR, Iozone work over Elan
6453 2002-08-07 Phil Schwan <phil@clusterfs.com>
6454 * version 0_5_3, our first alpha
6455 * we use the new Portals iovs
6456 * documentation updates
6457 * bug fixes and cleanups
6458 * small changes in the DLM wire protocol
6460 2002-07-25 Peter J. Braam <braam@clusterfs.com>
6461 * version 0_5_1 with some initial stability,
6462 * locking on MD and file I/O.
6463 * documentation updates
6464 * several bug fixes since 0.5.0
6465 * small changes in wire protocol
6467 2002-07-18 Phil Schwan <phil@clusterfs.com>
6469 * delivered as Lustre Light Alpha
6470 * fixed a crash after handling invalid MDS requests
6471 * fixed directory pages for architectures with non-4k pages sizes
6473 2002-07-11 Andreas Dilger <adilger@clusterfs.com>
6474 * release version v0_4_4
6475 * Moves TCP acceptor to be on port 2432 (unused Coda port) instead
6477 * Fixes a number of interruption problems with OST operations.
6478 * Update documentation for portals header changes
6479 * Move all wire protocol structs/defines to lustre_idl.h
6480 * Fixes symlink length bug.
6481 * Add tcpdump to repository.
6483 2002-07-05 Andreas Dilger <adilger@clusterfs.com>
6484 * release version v0_4_3
6485 * Fixes statfs for inodes on extN.
6486 * Fixes bug in runtests which would delete /etc/hosts.
6487 * Use 64-bit object IDs wherever possible (not into VFS though)
6488 Remove ost_get_info, which is unused by lustre, and out of date.
6490 2002-07-03 Peter Braam <braam@clusterfs.com>
6491 * release version v0_4_2 Fixes a lookup error (type not passed)
6492 * move forward to head of Portals
6493 * move forward to latest Lustre kernel
6495 2002-06-25 Peter Braam <braam@clusterfs.com>
6496 * release version v0_4_1. Hopefully stable on single node use.