1 tbd Sun Microsystems, Inc.
4 2.6.16.60-0.31 (SLES 10),
5 2.6.18-92.1.10.el5 (RHEL 5),
6 2.6.22.14 vanilla (kernel.org).
7 * Client support for unpatched kernels:
8 (see http://wiki.lustre.org/index.php?title=Patchless_Client)
9 2.6.16 - 2.6.21 vanilla (kernel.org)
10 * Recommended e2fsprogs version: 1.40.11-sun1
11 * Note that reiserfs quotas are disabled on SLES 10 in this kernel.
12 * RHEL 4 and RHEL 5/SLES 10 clients behaves differently on 'cd' to a
13 removed cwd "./" (refer to Bugzilla 14399).
14 * File join has been disabled in this release, refer to Bugzilla 16929.
17 Severity : enhancement
19 Description: Interoperability at server side (Disk interoperability)
21 Severity : enhancement
23 Description: Update to SLES10 SP2 kernel-2.6.16.60-0.31.
25 Severity : enhancement
27 Description: New client IO stack (CLIO).
29 Severity : enhancement
31 Description: Commit on sharing. Eliminate inter-client dependencies between
32 uncommitted transactions by doing transaction commits.
33 Thereby clients may recovery independently.
36 Frequency : Create a symlink file with a very long name
38 Description: ldlm_cancel_pack()) ASSERTION(max >= dlm->lock_count + count)
39 Details : If there is no extra space in the request for early cancels,
40 ldlm_req_handles_avail() returns 0 instead of a negative value.
42 Severity : enhancement
44 Description: Add /proc entry for import status
45 Details : The mdc, osc, and mgc import directories now have
46 an import directory that contains useful import data for debugging
49 Severity : enhancement
51 Description: Re-disable certain /proc logging
52 Details : Enable and disable client's offset_stats, extents_stats and
53 extents_stats_per_process stats logging on the fly.
56 Frequency : Only on FC kernels 2.6.22+
58 Description: oops in statahead
59 Details : Do not drop reference count for the dentry from VFS when lookup,
60 VFS will do that by itself.
62 Severity : enhancement
64 Description: Generic /proc file permissions
65 Details : Set /Proc file permissions in a more generic way to enable non-
66 root users operate on some /proc files.
70 Description: Hitting mdc_commit_close() ASSERTION
71 Details : Properly handle request reference release in
72 ll_release_openhandle().
76 Description: quota recovery deadlock during mds failover
77 Details : This patch includes att18982, att18236, att18237 in bz14840.
79 1. fix osts hang when mds does failover with quotaon
80 2. prevent watchdog storm when osts threads wait for the
85 Frequency : only patchless client
86 Description: add workaround for race between add/remove dentry from hash
88 Severity : enhancement
90 Description: Allow OST glimpses to return PW locks
94 Description: LBUG when llog conf file is full
95 Details : When llog bitmap is full, ENOSPC should be returned for plain log.
99 Description: Prevent import from entering FULL state when server in recovery
103 Description: service mount cannot take device name with ":"
104 Details : Only when device name contains ":/" will mount treat it as
110 Description: replace ptlrpcd with the statahead thread to interpret the async
111 statahead RPC callback
115 Frequency : on recovery
116 Description: I/O failures after umount during fail back
117 Details : if client reconnected to restarted server we need join to recovery
118 instead of find server handler is changed and process self
119 eviction with cancel all locks.
121 Severity : enhancement
123 Description: Update to RHEL5 kernel-2.6.18-92.1.10.el5.
125 Severity : enhancement
127 Description: Update to SLES10 SP2 kernel-2.6.16.60-0.27.
129 Severity : enhancement
131 Description: Upcall on Lustre log has been dumped
132 Details : Allow for a user mode script to be called once a Lustre log has
133 been dumped. It passes the filename of the dumped log to the
134 script, the location of the script can be specified via
135 /proc/sys/lnet/debug_log_upcall.
140 Description: avoid idr_remove called for id which is not allocated.
141 Details : Move assigment s_dev for clustered nfs to end of initialization,
142 to avoid problem with error handling.
147 Description: avoid Already found the key in hash [CONN_UNUSED_HASH] messages
148 Details : When connection is reused this not moved from CONN_UNUSED_HASH
149 into CONN_USED_HASH and this prodice warning when put connection
150 again in unused hash.
152 Severity : enhancement
154 Description: File striping can now be set to use an arbitrary pool of OSTs.
156 Severity : enhancement
158 Description: Export bytes_read/bytes_write count on OSC/OST.
162 Description: Early reply size mismatch, MGC loses connection
163 Details : Apply the MGS_CONNECT_SUPPORTED mask at reconnect time so
164 the connect flags are properly negotiated.
169 Description: quotas are not honored with O_DIRECT
170 Details : all writes with the flag O_DIRECT will use grants which leads to
171 this problem. Now using OBD_BRW_SYNC to guard this.
175 Description: add quota statistics
176 Details : 1. sort out quota proc entries and proc code.
177 2. add quota statistics
179 Severity : enhancement
181 Description: enable quota support for HEAD.
185 Description: Properly propagate oinfo flags from lov to osc for statfs
186 Details : restore missing copy oi_flags to lov requests.
188 Severity : enhancement
190 Description: Add man pages for llobdstat(8), llstat(8), plot-llstat(8),
191 : l_getgroups(8), lst(8), routerstat(8)
192 Details : included man pages for llobdstat(8), llstat(8),
193 : plot-llstat(8), l_getgroups(8), lst(8), routerstat(8)
195 Severity : enhancement
197 Description: configure's --enable-quota should check the
198 : kernel .config for CONFIG_QUOTA
199 Details : configure is terminated if --enable-quota is passed but
200 : no quota support is in kernel
204 Description: Remove portals compatibility
205 Details : Remove portals compatibility, not interoperable with releases
210 Description: Resolve device initialization race
211 Details : Prevent proc handler from accessing devices added to the
212 obd_devs array but yet be intialized.
214 Severity : enhancement
216 Description: Update to SLES10 SP2 kernel-2.6.16.60-0.23.
218 Severity : enhancement
220 Description: Update to RHEL5 kernel-2.6.18-92.1.6.el5.
225 Description: Using wrong pointer in osc_brw_prep_request
226 Details : Access to array[-1] can produce panic if kernel compiled with
227 CONFIG_PAGE_ALLOC enabled
231 Description: Client runs out of low memory
232 Details : Consider only lowmem when counting initial number of llap pages
236 Description: *optional* service tags registration
237 Details : if the "service tags" package is installed on a Lustre node
238 When the filesystem is mounted, a local-node service tag will
239 be created. See http://inventory.sun.com/ for more information
240 about the Service Tags asset management system.
244 Description: Kernel BUG tries to release flock
245 Details : Lustre does not destroy flock lock before last reference goes
246 away. So always drop flock locks when client is evicted and
247 perform unlock regardless of successfulness of speaking to MDS.
251 Description: add refcount protection for osc callbacks, avoid panic on shutdown
255 Description: sanity test 65a fails if stripecount of -1 is set
256 Details : handle -1 striping on filesystem in ll_dirstripe_verify
261 Description: ASSERTION(CheckWriteback(page,cmd)) failed
262 Details : badly clear PG_Writeback bit in ll_ap_completion can produce false
265 Severity : enhancement
267 Description: Update to RHEL5 kernel-2.6.18-53.1.21.el5.
271 Description: do not process already freed flock
272 Details : flock can possibly be freed by another thread before it reaches
273 to ldlm_flock_completion_ast.
277 Description: LBUG during stress test
278 Details : Need properly lock accesses the flock deadlock detection list.
282 Description: oops in page fault handler
283 Details : kernel page fault handler can return two special 'pages' in error
284 case, don't try dereference NOPAGE_SIGBUS and NOPAGE_OMM.
288 Description: timeout with invalidate import.
289 Details : ptlrpcd_check call obd_zombie_impexp_cull and wait request which
290 should be handled by ptlrpcd. This produce long age waiting and
291 -ETIMEOUT ptlrpc_invalidate_import and as result LASSERT.
293 Severity : enhancement
295 Description: Update to RHEL5 kernel-2.6.18-53.1.19.el5.
299 Description: enable MGS and MDT services start separately
300 Details : add a 'nomgs' option in mount.lustre to enable start a MDT with
301 a co-located MGS without starting the MGS, which is a complement
302 to 'nosvc' mount option.
306 Frequency : after recovery
307 Description: precreate to many object's after del orphan.
308 Details : del orphan st in oscc last_id == next_id and this triger growing
309 count of precreated objects. Set flag LOW to skip increase count
310 of precreated objects.
314 Frequency : rare, on clear nid stats
315 Description: ASSERTION(client_stat->nid_exp_ref_count == 0)
316 Details : when clean nid stats sometimes try destroy live entry,
317 and this produce panic in free.
321 Description: Stack overflow during MDS log replay
322 ease stack pressure by using a thread dealing llog_process.
326 Description: wait until IO finished before start new when do lock cancel.
327 Details : VM protocol want old IO finished before start new, in this case
328 need wait until PG_writeback is cleared until check dirty flag and
329 call writepages in lock cancel callback.
331 Severity : enhancement
333 Description: using special macro for print time and cleanup in includes.
337 Description: mds_mfd_close() ASSERTION(rc == 0)
338 Details : In mds_mfd_close(), we need protect inode's writecount change
339 within its orphan write semaphore to prevent possible races.
343 Description: Obsolete CURRENT_SECONDS and use cfs_time_current_sec() instead.
347 Frequency : rare, on shutdown ost
348 Description: don't hit live lock with umount ost.
349 Details : shrink_dcache_parent can be in long loop with destroy dentries,
350 use shrink_dcache_sb instead.
354 Description: don't panic with use echo client
355 Details : echo client pass NULL as client nid pointer and this produce null
360 Description: fix build on ppc32
361 Details : compile code with -m64 flag produce wrong object file for ppc32.
365 Description: add message levels for liblustreapi
369 Description: fix for occasional failure case of -ENOSPC in recovery-small tests
370 Details : Move the 'good_osts' check before the 'total_bavail' check. This
371 will result in an -EAGAIN and in the exit call path we call
372 alloc_rr() which will with increasing aggressiveness attempt to
373 aquire precreated objects on the minimum number of required OSCs.
377 Description: Use old size assignment to avoid deadlock
378 Details : Reverts the changes in bugs 2369 and bug 14138 that introduced
379 the scheduling while holding a spinlock. We do not need locking
380 for size in ll_update_inode() because size is only updated from
381 the MDS for directories or files without objects, so there is no
382 other place to do the update, and concurrent access to such inodes
383 are protected by the inode lock.
387 Description: resolve "_IOWR redefined" build error on SLES10
391 Description: dump the memory debugging after all modules are unloaded to
392 suppress false negative in conf_sanity test 39
394 Severity : enhancement
396 Description: build kernel-ib packages for OFED 1.3 in our release cycle
401 Description: fix SLES kernel versioning
402 Details : the kernel version for our SLES 10 kernel did not include a "-"
403 before the "smp" at the end. while this was not a problem in
404 general, it did mean that software trying to use the kernel
405 version to try to detect a vendor specific kernel would fail.
406 this was most evident by the OFED build scripts.
410 Description: Don't update lov_desc members until making sure they are valid
411 Details : When updating lov_desc members via proc fs, need fix their
412 validities before doing the real update.
416 Description: don't put request into delay list while invalidate in flight.
417 Details : ptlrpc_delay_request sometimes put in delay list while invalidate
418 import in flight. this produce timeout for invalidate and sometimes
419 can cause stale data.
423 Frequency : on ppc only
424 Description: not convert ost objects for directory because it's not exist.
425 Details : ll_dir_getstripe assume dirrectory has ost objects but this wrong.
429 Description: Add FMODE_EXEC file flag for SLES10 SP1 kernel.
431 Severity : enhancement
433 Description: Update to support 2.6.22.14 vanilla kernel.
437 Frequency : rare, on recovery
438 Description: read procfs can produce deadlock in some situation
439 Details : Holding lprocfs lock which send rpc can produce block for destroy
440 obd objects and this also block reconnect with -EALREADY. This
441 isn't fix all lprocfs bugs - but make it rare.
443 Severity : enhancement
445 Description: Update kernel to RHEL5 2.6.18-53.1.14.el5.
448 Frequency : frequent on X2 node
450 Description: mdc_set_open_replay_data LBUG
451 Details : Set replay data for requests that are eligible for replay.
455 Description: lustre_mgs: operation 101 on unconnected MGS
456 Details : When MGC is disconnected from MGS long enough, MGS will evict the
457 MGC, and late on MGC cannot successfully connect to MGS and a lot
458 of the error messages complaining that MGS is not connected.
462 Frequency : on network error
463 Description: panic with double free request if network error
464 Details : mdc_finish_enqueue is finish request if any network error ocuring,
465 but it's true only for synchronus enqueue, for async enqueue
466 (via ptlrpcd) this incorrect and ptlrpcd want finish request
469 Severity : enhancement
471 Description: client-side metadata stat-ahead during readdir(directory readahead)
472 Details : perform client-side metadata stat-ahead when the client detects
473 readdir and sequential stat of dir entries therein
476 Frequency : on start mds
478 Description: Implement get_info(last_id) in obdfilter.
481 Frequency : occasional
483 Description: Correctly check stale fid, not start epoch if ost not support SOM
484 Details : open with flag O_CREATE need set old fid in op_fid3 because
485 op_fid2 was overwritten with new generated fid, but mds can answer
486 with one of these two fids and both is not stale. Setattr
487 incorrectly started an epoch and assume will be called
488 done_writing, but without SOM done_writing ever being called.
491 Frequency : rare, depends on device drivers and load
493 Description: MDS or OSS nodes crash due to stack overflow
494 Details : Code changes in 1.8.0 increased the stack usage of some functions.
495 In some cases, in conjunction with device drivers that use a lot
496 of stack the MDS (or possibly OSS) service threads could overflow
497 the stack. One change which was identified to consume additional
498 stack has been reworked to avoid the extra stack usage.
501 Frequency : occasional
503 Description: Do not fail import if osc_interpret_create gets -EAGAIN
504 Details : If osc_interpret_create got -EAGAIN it immediately exits and
505 wakeup oscc_waitq. After wakeup oscc_wait_for_objects call
506 oscc_has_objects and see OSC has no objests and call
507 oscc_internal_create to resend create request.
509 Severity : enhancement
511 Description: Update to SLES10 SP1 latest kernel-2.6.16.54-0.2.5.
513 Severity : enhancement
515 Description: Update to RHEL5 latest kernel-2.6.18-53.1.13.el5.
518 Frequency : very rare
520 Description: Fix replay if there is an un-replied request and open
521 Details : In some cases, older replay request will revert the
522 mcd->mcd_last_xid on MDS which is used to record the client's
525 Severity : enhancement
527 Description: Update to RHEL5 latest kernel-2.6.18-53.1.6.el5.
529 Severity : enhancement
531 Description: Add rhel5 support to HEAD.
533 Serverity : enhancement
535 Description: Update RHEL4 kernel to 2.6.9-67.0.4.
540 Description: Don't allow skipping OSTs if index has been specified.
541 Details : Don't allow skipping OSTs if index has been specified, make
542 locking in internal create lots better.
546 Description: LBUG in ptlrpc_check_set() bad phase ebc0de00
547 Details : access to bitfield in structure is always rounded to long
548 and this produce problem with not atomic change any bit.
552 Description: Lustre make rpms failed.
553 Details : Remove ldiskfs spec file to avoids rpmbuild be confused when
554 builds Lustre rpms from tarball.
557 Frequency : rare on shutdown ost
559 Description: If llog cancel was not send before clean_exports phase, this can
560 produce deadlock in llog code.
561 Details : If llog thread has last reference to obd and call class_import_put
562 this produce deadlock because llog_cleanup_commit_master wait when
563 last llog_commit_thread exited, but this never success because was
564 called from llog_commit_thread.
568 Description: allow userland application know is lost one of stripes.
569 Details : fill lvb_blocks with error code on ost and return it to
570 application if error flag found.
574 Description: NULL lov_tgts causing MDS oops
575 Details : more safe checks for NULL lov_tgts for avoid oops.
577 Severity : enhancement
579 Description: Update to RHEL4 latest kernel-2.6.9-67.0.1.EL.
583 Descriptoin: make lov_create() will not stuck in obd_statfs_rqset()
584 Details : If an OST is down the MDS will hang indefinitely in
585 obd_statfs_rqset() waiting for the statfs data. While for
586 MDS QOS usage of statfs, it should not stuck in waiting.
588 Severity : enhancement
590 Description: remote_acl support
591 Details : Support ACL-based permission check for remote user.
592 Support setfacl/getfacl for remote user with the utils
593 "lfs {l,r}{s,g}etfacl" which follow the same parameter format as
594 the system "{s,g}etfacl" utils.
596 Severity : enhancement
598 Description: Update to RHEL4 U6 kernel-2.6.9-67.EL.
600 Severity : enhancement
602 Description: Update to RHEL5 latest kernel-2.6.18-53.1.4.el5.
606 Description: make mgs_setparam() handle fsname containing dash
607 Details : fsname containing a dash does not work with lctl conf_param
609 Severity : enhancement
611 Description: Update to SLES10 SP1 latest kernel-2.6.16.54-0.2.3.
613 Severity : enhancement
615 Description: Update to RHEL5 Update-1 kernel 2.6.18-53.el5.
619 Frequency : rare, at shutdown
620 Description: access already free / zero obd_namespace.
621 Details : if client_disconnect_export was called without force flag set,
622 and exist connect request in flight, this can produce access to
623 NULL pointer (or already free pointer) when connect_interpret
624 store ocd flags in obd_namespace.
628 Frequency : only at startup
629 Description: not alloc memory with spinlock held.
630 Details : allocation memory with GFP_KERNEL can produce sleep deadlock,
631 if any spinlock held.
633 Severity : enhancement
635 Description: make lustre randomly fail allocating memory
636 Details : Make lustre randomly failed allocating memory for testing purpose.
638 Severity : enhancement
640 Description: lost problems with lov objid file
641 Details : Fixes some scability and access to not inited memory problems
642 in work with lov objdid file.
647 Description: lfs find does not continue on file error
648 Details : Continue other files processing when a file/dir is absent.
652 Description: Inconsistent usage of lustre_pack_reply()
653 Details : Standardize the usage of lustre_pack_reply() such that it
654 always generate a CERROR on failure.
657 Frequency : occasional
659 Description: MDS hang or stay in waiting lock
660 Details : If client receive lock with CBPENDING flag ldlm need send lock
661 cancel as separate rpc, to avoid situation when cancel request
662 can't processed due all i/o threads stay in wait lock.
666 Description: Update to RHEL5 kernel 2.6.18-8.1.15.el5.
670 Description: Update to SLES10 SP1 kernel 2.6.16.53-0.16
674 Description: Update to SLES9 kernel-2.6.5-7.287.3.
678 Description: Update to RHEL4 latest kernel.
680 Severity : enhancement
682 Description: Build SLES10 patchless client fails
683 Details : The configure was broken by run ./configure with
684 --with-linux-obj=.... argument for patchless client. When the
685 configure use --with-linux-obj, the LINUXINCLUDE= -Iinclude
686 can't search header adequately. Use absolute path such as
687 -I($LINUX)/include instead.
691 Description: interrupt oig_wait produce painc on resend.
692 Details : brw_redo_request can be used for resend requests from ptlrpcd and
693 private set, and this produce situation when rq_ptlrpcd_data not
694 copyed to new allocated request and triggered LBUG on assert
695 req->rq_ptlrpcd_data != NULL. But this member used only for wakeup
696 ptlrpcd set if request is changed and can be safety changed to use
701 Description: LASSERT_{REQ,REP}SWAB macros are buggy
702 Details : If SWAB_PARANOIA is disabled, the LASSERT_REQSWAB and
703 LASSERT_REPSWAB macros become no-ops, which is incorrect. Drop
704 these macros and replace them with their difinitions instead.
708 Description: Update kernel patches for SLES10 2.6.16.53-0.8.
709 Details : Update which_patch & target file for SLES10 latest kernel.
712 Description: Remove client patches from SLES 10 kernel.
713 Details : This causes SLES 10 clients to behave as patchless clients
714 even on a Lustre-patched (server) kernel.
716 Severity : enhancement
718 Description: self-adjustable client's lru lists
719 Details : use adaptive algorithm for managing client cached locks lru
720 lists according to current server load, other client's work
721 pattern, memory activities, etc. Both, server and client
722 side namespaces provide number of proc tunables for controlling
725 Severity : enhancement
727 Description: light-weight GSS support
728 Details : Support krb5n and krb5a mode, which keep Kerberos 5 authentication
729 and reduce performance overhead.
731 Severity : enhancement
733 Description: Linux keyring support
734 Details : Support using service of Linux keyring for Lustre GSS internal
735 context refresh/cache mechanism.
739 Description: Fix errors in lfs documentation
740 Details : Fixes man pages
744 Description: don't use GFP_* in generic Lustre code.
745 Details : Use cfs_alloc_* functions and CFS_* flags for code portability.
749 Description: obdclass is limited by single OBD_ALLOC(idarray)
750 Details : replace OBD_ALLOC/OBD_FREE with OBD_VMALLOC/OBD_VFREE
754 Description: warnings with build patchless client with vanila 2.6.19 and up
755 Details : change old ctl_table style and replace ctl_table/ctl_table_header
756 with cfs_sysctl_table_t/cfs_sysctl_table_header_t
760 Frequency : Only for SLES
761 Description: sanity_quota fail test_1
762 Details : There are multiple occurences of $TSTUSR in SLES's /etc/group
763 file, which makes TSTID[2] inunique.
767 Frequency : Only for SLES9
768 Description: Kernel patches for SLES9 2.6.5-7.286 kernel
769 Details : Update target/ChangeLog/which_patch .
773 Frequency : Only for test_10 in sanity-quota.sh
774 Description: a bug in quota test code
775 Details : checking a test flag in a wrong place
779 Frequency : Only for test_13 in sanity-quota.sh
780 Description: a bug in quota test script
781 Details : shouldn't check overfull allocated quota
785 Description: jbd statistics
786 Details : Port older jbd statistics patch for sles10
788 Severity : enhancement
790 Description: Add printing inode info into message about error in writepage.
795 Description: reply_lock_interpret crash due to race with it and lock cancel.
796 Details : Do not replay locks that are being cancelled. Do not reference
797 locks by their address during replay, just by their handle.
801 Frequency : When flocks are used.
802 Description: assertion failure in ldlm_cli_enquque_fini for non NULL lock.
803 Details : Flock locks might destroy just granted lock if it could be merged
804 with another existing flock, this is done in completion handler,
805 so teach ldlm_cli_enquque_fini that this is a valid case for
811 Description: Oops in read and write path when failing to allocate lock.
812 Details : Check if lock allocation failed and return error back.
816 Description: lstripe command fails for valid OST index
817 Details : The stripe offset is compared to lov->desc.ld_tgt_count
818 instead of lov->desc.ld_active_tgt_count.
822 Description: sanity.sh failed test 103
823 Details : RHEL mis-interpret setfacl "-X" param, so we won't test setfacl
828 Description: df doesn't work properly if diskfs blocksize != 4K
829 Details : Choose biggest blocksize of OST's as the LOV's blocksize.
833 Description: lfs find on -1 stripe looping in lsm_lmm_verify_common()
834 Details : Avoid lov_verify_lmm_common() on directory with -1 stripe count.
836 Severity : enhancement
838 Description: Adaptive timeouts
839 Details : RPC timeouts adapt to changing server load and network
840 conditions to reduce resend attempts and improve recovery time.
844 Description: llapi_file_create() does not allow some changes
845 Details : add llapi_file_open() that allows specifying the mode and
846 open flags, and also returns an open file handle.
850 Description: merge and cleanup kernel patches.
851 Details : Remove mnt_lustre_list in vfs_intent-2.6-rhel4.patch.
855 Description: Add journal checksum support.(Kernel part)
856 Details : The journal checksum feature adds two new flags i.e
857 JBD2_FEATURE_INCOMPAT_ASYNC_COMMIT and
858 JBD2_FEATURE_COMPAT_CHECKSUM. JBD2_FEATURE_CHECKSUM flag
859 indicates that the commit block contains the checksum for
860 the blocks described by the descriptor blocks. Now commit
861 record can be sent to disk without waiting for descriptor
862 blocks to be written to disk. This behavior is controlled
863 using JBD2_FEATURE_ASYNC_COMMIT flag.
867 Description: OSS needs mutliple precreate threads
868 Details : Add ability to start more than one create thread per OSS.
872 Description: Sanity.sh test_65k failed: setstripe should have succeeded
873 Details : execute lfs setstripe on client
877 Description: mds_obd_create error creating tmp object
878 Details : When the user sets quota on root, llog will be affected and can't
879 create files and write files.
882 Frequency : Always on ia64 patchless client, and possibly others.
884 Description: Add EXPORT_SYMBOL check for node_to_cpumask symbol.
885 Details : This allows the patchless client to be loaded on architectures
890 Description: RedHat Update kernel for RHEL5
891 Details : Add the kernel config file for RHEL5.
895 Description: RedHat Update kernel for RHEL5
896 Details : Modify the kernel config file more closer RHEL5.
900 Description: Build failure against Centos5 (RHEL5)
901 Details : Define PAGE_SIZE when it isn't present.
905 Description: test_7 on sanity-quota.sh can't be used on separate machines
906 Details : add facet to handle it
910 Description: "ll_intent_file_open()) lock enqueue: err: -13" with nfs
911 Details : with NFS, the anon dentry's parent was set to itself in
912 d_alloc_anon(), so in MDS, we use rec->ur_fid1 to find the
913 corresponding dentry other than use rec->ur_name.
915 Severity : enhancement
917 Description: lfs setstripe enhancement
918 Details : Make lfs setstripe understand 'k', 'm' and 'g' for stripe size.
922 Description: enable data checksumming by default
923 Details : enable checksum by default, allow --disable-checksum
924 configure option and "-o nochecksum" mount option. Checksums
925 can also be disabled at runtime via $LPROC/osc/*/checksum_pages.
929 Description: lustre support for RHEL5
930 Details : Add support for RHEL5.
934 Description: Client eviction due to failover config
935 Details : after a connection loss, the lustre client should attempt to
936 reconnect to the last active server first before trying the
937 other potential connections.
941 Description: when mds and osts use different quota unit(32bit and 64bit),
942 quota will be released repeatly.
943 Details : void sending multiple quota reqs to mds, which will keep the
944 status between the reqs.
948 Description: osts not allocated evenly to files
949 Details : change the condition to increase offset_idx
952 Frequency : Always for filesystems larger than 2TB on 32-bit systems.
954 Description: Data corruption for OSTs that are formatted larger than 2TB
956 Details : When generating the bio request for lustre file writes the
957 sector number would overflow a temporary variable before being
958 used for the IO. The data reads correctly from Lustre (which
959 will overflow in a similar manner) but other file data or
960 filesystem metadata may be corrupted in some cases.
964 Description: Tune the kernel for good SCSI performance.
965 Details : Set the value of /sys/block/{dev}/queue/max_sectors_kb
966 to the value of /sys/block/{dev}/queue/max_hw_sectors_kb
971 Description: rewrite ext2-derived code in obdclass/uuid.c
972 Details : rewrite inherited code (uuid parsing code from ext2 utils)
973 from scratch preserving functionality.
977 Description: "lfs find -obd UUID" prints directories
978 Details : "lfs find -obd UUID" will return all directory names instead
979 of just file names. It is incorrect because the directories
980 do not reside on the OSTs.
984 Description: use i_size_read and i_size_write in 2.6 port
985 Details : replace inode->i_size access with i_size_read/write()
988 Frequency : when removing large files
990 Description: scheduling issue during removal of large Lustre files
991 Details : Don't take the BKL in fsfilt_ext3_setattr() for 2.6 kernels.
992 It causes scheduling issues when removing large files (17TB in the
996 Frequency : only with liblustre clients on XT3
998 Description: evictions taking too long
999 Details : allow llrd to evict clients directly on OSTs
1002 Frequency : only on ppc
1004 Description: /proc/fs/lustre/devices broken on ppc
1005 Details : The patch as applied to 1.6.2 doesn't look correct for all arches.
1006 We should make sure the type of 'index' is loff_t and then cast
1007 explicitly as needed below. Do not assign an explicitly cast
1011 Frequency : only for rhel5
1013 Description: Kernel patches update for RHEL5 2.6.18-8.1.10.el5.
1014 Details : Modify the target file & which_kernel.
1016 Severity : enhancement
1018 Description: omit set fsid for export NFS
1019 Details : fix set/restore device id for avoid EMFILE error and mark lustre
1020 fs as FS_REQUIRES_DEV for avoid problems with generate fsid.
1024 Frequency : Always, for kernels after 2.6.16
1025 Description: Fix warning idr_remove called for id=.. which is not allocated.
1026 Details : Last kernels save old s_dev before kill super and not allow
1027 to restore from callback - restore it before call kill_anon_super.
1031 Description: buffer overruns could theoretically occur
1032 Details : llapi_semantic_traverse() modifies the "path" argument by
1033 appending values to the end of the origin string, and a
1034 overrun may occur. Adding buffer overrun check in liblustreapi.
1038 Description: Fix error on 'ls .' at the top of the Lustre mount.
1039 Details : Don't revalidate dentry if it is a root dentry.
1043 Description: Kernel patches update for RHEL4 2.6.9-55.0.6.
1044 Details : Modify vm-tunables-rhel4.patch.
1048 Description: Kernel config for 2.6.18-vanilla.
1049 Details : Modify targets/2.6-vanilla.target.in.
1050 Add config file kernel-2.6.18-2.6-vanilla-i686.config.
1051 Add config file kernel-2.6.18-2.6-vanilla-i686-smp.config.
1052 Add config file kernel-2.6.18-2.6-vanilla-x86_64.config.
1053 Add config file kernel-2.6.18-2.6-vanilla-x86_64-smp.config.
1057 Description: improve handling recoverable errors
1058 Details : if request processig with error which can be recoverable on server
1059 request should be resend, otherwise page released from cache and
1064 Description: Kernel patches update for RHEL5 2.6.18-8.1.14.el5.
1065 Details : Modify target file & which_patch.
1066 A flaw was found in the IA32 system call emulation provided
1067 on AMD64 and Intel 64 platforms. An improperly validated 64-bit
1068 value could be stored in the %RAX register, which could trigger an
1069 out-of-bounds system call table access. An untrusted local user
1070 could exploit this flaw to run code in the kernel
1071 (ie a root privilege escalation). (CVE-2007-4573).
1075 Description: O_DIRECT bypasses client statistics.
1076 Details : When running with O_DIRECT I/O, neither the client rpc_stats nor
1077 read_ahead_stats were updated. copied the stats section from
1078 osc_send_oap_rpc() into async_internal().
1082 Description: Add jbd statistics patch for RHEL5 and 2.6.18-vanilla
1086 Description: change order of libsysio includes
1087 Details : '#include sysio.h' should always come before '#include xtio.h'
1091 Description: handle "serious error: objid * already exists" more gracefully
1092 Details : If LAST_ID value on disk is smaller than the objects existing in
1093 the O/0/d* directories, it indicates disk corruption and causes an
1094 LBUG(). If the object is 0-length, then we should use the existing
1095 object. This will help to avoid a full fsck in most cases.
1097 Severity : enhancement
1099 Description: adapt the lustre_config script to support the upgrade case
1100 Details : Add "-u" option for lustre_config script to support upgrading 1.4
1101 server targets to 1.6 in parallel.
1105 Description: To avoid grant space > avaible space when the disk is almost
1106 full. Without this patch you might see the error "grant XXXX >
1107 available" or some LBUG about grant, when the disk is almost
1109 Details : In filter_check_grant, for non_grant cache write, we should
1110 check the left space by if (*left > ungranted + bytes), instead
1111 of (*left > ungranted), because only we are sure the left space
1112 is enough for another "bytes", then the ungrant space should be
1113 increase. In client, we should update cl_avail_grant only there
1114 is OBD_MD_FLGRANT in the reply.
1118 Description: Update RHEL 4 kernel to fix local root privilege escalation.
1119 Details : Update to the latest RHEL 4 kernel to fix the vulnerability
1120 described in CVE-2007-4573. This problem could allow untrusted
1121 local users to gain root access.
1124 Frequency : when using O_DIRECT and quotas
1126 Description: Incorrect file ownership on O_DIRECT output files
1127 Details : block usage reported by 'lfs quota' does not take into account
1128 files that have been written with O_DIRECT.
1133 Description: touch file failed when fs is not full
1134 Details : OST in recovery should not be discarded by MDS in alloc_qos(),
1135 otherwise we can get ENOSP while fs is not full.
1139 Description: parallel lock callbacks
1140 Details : Instead of sending blocking and completion callbacks as separated
1141 requests, adding them to a set and sending in parallel.
1144 Frequency : only for Cray XT3
1145 Bugzilla : 12829/13455
1146 Description: Changing primary group doesn't change the group lustre assigns to
1148 Details : When CRAY_XT3 is defined, the fsgid supplied by the client is
1149 overridden with the primary group provided by the group upcall,
1150 whereas the supplied fsgid can be trusted if it is in the list of
1151 supplementary groups returned by the group upcall.
1153 Severity : enhancement
1155 Description: Allow masking D_WARNING, D_ERROR messages from console
1156 Details : Console messages can now be disabled via lnet.printk.
1160 Description: User code with malformed file open parameter crashes client node
1161 Details : Before packing join_file req, all the related reference should be
1162 checked carefully in case some malformed flags cause fake
1163 join_file req on client.
1167 Description: LDLM_ENQUEUE races with LDLM_CP_CALLBACK
1168 Details : ldlm_completion_ast() assumes that a lock is granted when the req
1169 mode is equal to the granted mode. However, it should also check
1170 that LDLM_FL_CP_REQD is not set.
1174 Description: Heavy nfs access might result in deadlocks
1175 Details : After ELC code landed, it is now improper to enqueue any mds
1176 locks under och_sem, because enqueue might want to decide to
1177 cancel open locks for same inode we are holding och_sem for.
1181 Description: Client eviction while running blogbench
1182 Details : A lot of unlink operations with concurrent I/O can lead to a
1183 deadlock causing evictions. To address the problem, the number of
1184 oustanding OST_DESTROY requests is now throttled to
1185 max_rpcs_in_flight per OSC and LDLM_FL_DISCARD_DATA blocking
1186 callbacks are processed in priority.
1190 Description: enable ACLs on MDS by default
1191 Details : ACLs must be enabled on MDS by default.
1194 Frequency : PPC/PPC64 only
1196 Description: conflicts between asm-ppc64/types.h and lustre_types.h
1197 Details : fix duplicated definitions between asm-ppc64/types.h and
1198 lustre_types.h on PPC.
1201 Frequency : PPC/PPC64 only
1203 Description: asm-ppc/segment.h does not exist
1204 Details : fix compile issue on PPC.
1208 Description: better handle error messages in extents code
1211 Frequency : RHEL4 only
1213 Description: mkfs is very slow on IA64/RHEL4
1214 Details : A performance regression has been discovered in the MPT Fusion
1215 driver between versions 3.02.73rh and 3.02.99.00rh. As a
1216 consequence, we have downgraded the MPT Fusion driver in the RHEL4
1217 kernel from 3.02.99.00 to 3.02.73 until this problem is fixed.
1219 Severity : enhancement
1221 Description: SNMP support enhancement
1222 Details : Adding total number of sampled request for an MDS node in snmp
1225 Severity : enhancement
1227 Description: Optimize ldlm waiting list processing for PR extent locks
1228 Details : When processing waiting list for read extent lock and meeting read
1229 lock that is same or wider to it that is not contended, skip
1230 processing rest of the list and immediatelly return current
1231 status of conflictness, since we are guaranteed there are no
1232 conflicting locks in the rest of the list.
1236 Description: Time out and refuse to reconnect
1237 Details : When the failover node is the primary node, it is possible
1238 to have two identical connections in imp_conn_list. We must
1239 compare not conn's pointers but NIDs, otherwise we can defeat
1240 connection throttling.
1244 Description: port llog fixes from b1_6 into HEAD
1245 Details : Port llog reference couting and some llog cleanups from b1_6
1246 (bug 10800) into HEAD, for protect from panic and access to already
1247 free llog structures.
1251 Description: Detect stride IO mode in read-ahead
1252 Details : When a client does stride read, read-ahead should detect that and
1253 read-ahead pages according to the detected stride pattern.
1257 Description: data checksumming impacts single node performance
1258 Details : add support for several checksum algorithm. Currently, only CRC32
1259 and Adler-32 are supported. The checksum type can be changed on
1260 the fly via /proc/fs/lustre/osc/*/checksum_type.
1264 Description: use adler32 for page checksums
1265 Details : when available, use the Adler-32 algorithm instead of CRC32 for
1270 Description: build for x2 fails
1271 Details : fix compile issue on Cray systems.
1275 Description: Properly match for duplicate locks
1276 Details : Due to different lock order from skiplists code, we need to
1277 traverse entire list for now
1280 Frequency : only on PPC/SLES10
1282 Description: "BITS_PER_LONG is not 32 or 64" in linux/idr.h
1283 Details : On SLES10/PPC, fs.h includes idr.h which requires BITS_PER_LONG to
1284 be defined. Add a hack in mkfs_lustre.c to work around this compile
1289 Description: LASSERT on MDS when client holding flock lock dies
1290 Details : ldlm pool logic depends on number of granted locks equal to
1291 number of released locks which is not true for flock locks, so
1292 just exclude such locks from consideration.
1296 Description: MDS deadlock with many ll_sync_lov threads and I/O stalled
1297 Details : Use fsfilt_sync() for both the whole filesystem sync and
1298 individual file sync to eliminate dangerous inode locking
1299 with I_LOCK that can lead to a deadlock.
1303 Description: performance in 1.6.3
1304 Details : Force q->max_phys_segments to MAX_PHYS_SEGMENTS on SLES10 to be
1305 sure that 1MB requests are not fragmented by the block layer.
1307 Severity : enhancement
1309 Description: organize the server-side client stats on per-nid basis
1310 Details : Change the structure of stats under obdfilter and mds to
1318 The "uuid"s file would list the uuids of _active_ exports.
1319 And the clear entry is to clear all stats and stale nids.
1321 Severity : enhancement
1323 Description: eliminate client locks in face of contention
1324 Details : file contention detection and lockless i/o implementation
1325 for contended files.
1329 Description: Reinitialize optind to 0 so that interactive lfs works in all cases
1332 Frequency : very rare, if additional xattrs are used on kernels >= 2.6.12
1334 Description: MDS may lose file striping (and hence file data) in some cases
1335 Details : If there are additional extended attributes stored on the MDS,
1336 in particular ACLs, SELinux, or user attributes (if user_xattr
1337 is specified for the client mount options) then there is a risk
1338 of attribute loss. Additionally, the Lustre file striping
1339 needs to be larger than default (e.g. striped over all OSTs),
1340 and an additional attribute must be stored initially in the
1341 inode and then increase in size enough to be moved to the
1342 external attribute block (e.g. ACL growing in size) for file
1347 Description: skiplist implementation simplification
1348 Details : skiplists are used to group compatible locks on granted list
1349 that was implemented as tracking first and last lock of each
1350 lock group the patch changes that to using doubly linked lists
1354 Description: MDS LBUG: ASSERTION(!IS_ERR(dchild))
1355 Details : Change LASSERTs to client eviction (i.e. abort client's recovery)
1356 because LASSERT on both the data supplied by a client, and the
1357 data on disk is dangerous and incorrect.
1359 Severity : enhancement
1361 Description: Slow truncate/writes to huge files at high offsets.
1362 Details : Directly associate cached pages to lock that protect those pages,
1363 this allows us to quickly find what pages to write and remove
1364 once lock callback is received.
1368 Description: more ldlm soft lockups
1369 Details : In ldlm_resource_add_lock(), call to ldlm_resource_dump()
1370 starve other threads from the resource lock for a long time in
1371 case of long waiting queue, so change the debug level from
1372 D_OTHER to the less frequently used D_INFO.
1374 Severity : enhancement
1376 Description: add -gid, -group, -uid, -user options to lfs find
1380 Description: Hung threads in invalidate_inode_pages2_range
1381 Details : The direct IO path doesn't call check_rpcs to submit a new RPC
1382 once one is completed. As a result, some RPCs are stuck in the
1383 queue and are never sent.
1387 Description: filter threads hungs on waiting journal commit
1388 Details : Cleanup filter group llog code, then only filter group llog will
1389 be only created in the MDS/OST syncing process.
1393 Description: Procfs and llog threads access destoryed import sometimes.
1394 Details : Sync the import destoryed process with procfs and llog threads by
1395 the import refcount and semaphore.
1397 Severity : enhancement
1399 Description: openlock cache of b1_6 port to HEAD
1404 Description: kernel BUG at ldiskfs2_ext_new_extent_cb
1405 Details : If insertion of an extent fails, then discard the inode
1406 preallocation and free data blocks else it can lead to duplicate
1411 Description: don't always update ctime in ext3_xattr_set_handle()
1412 Details : Current xattr code updates inode ctime in ext3_xattr_set_handle.
1413 In some cases the ctime should not be updated, for example for
1414 2.0->1.8 compatibility it is necessary to delete an xattr and it
1415 should not update the ctime.
1419 Bugzilla : 15713/16362
1420 Description: Assertion in iopen_connect_dentry in 1.6.3
1421 Details : looking up an inode via iopen with the wrong generation number can
1422 populate the dcache with a disconneced dentry while the inode
1423 number is in the process of being reallocated. This causes an
1424 assertion failure in iopen since the inode's dentry list contains
1425 both a connected and disconnected dentry.
1429 Description: assertion failure in ldlm_handle2lock()
1430 Details : fix a race between class_handle_unhash() and class_handle2object()
1431 introduced in lustre 1.6.5 by bug 13622.
1436 Description: Kernel BUG: sd_iostats_bump: unexpected disk index
1437 Details : remove the limit of 256 scsi disks in the sd_iostat patch
1442 Description: oops in sd_iostats_seq_show()
1443 Details : unloading/reloading the scsi low level driver triggers a kernel
1444 bug when trying to access the sd iostat file.
1449 Description: Kernel panics during QLogic driver reload
1450 Details : REQ_BLOCK_PC requests are not handled properly in the sd iostat
1451 patch, causing memory corruption.
1456 Description: journal_dev option does not work in b1_6
1457 Details : pass mount option during pre-mount.
1459 Severity : enhancement
1461 Description: Add a FIEMAP(FIle Extent MAP) ioctl
1462 Details : FIEMAP ioctl will allow an application to efficiently fetch the
1463 extent information of a file. It can be used to map logical blocks
1464 in a file to physical blocks in the block device.
1468 Description: LDLM soft lockups - improvement
1469 Details : It is be possible to send the lock handle along with each read
1470 or write request because the client is already doing a lock match
1471 itself so there isn't any reason the OST should have to re-do that
1475 Frequency : only on Cray X2
1477 Description: X2 build failures
1478 Details : fix build failures on Cray X2.
1482 Description: xid & resent requests
1483 Details : Initialize RPC XID from clock at startup (randomly if clock is
1486 Severity : enhancement
1488 Description: Add lustre_start utility to start or stop multiple Lustre servers
1493 Description: Lustre GPF in {:ptlrpc:ptlrpc_server_free_request+373}
1494 Details : In case of memory pressure, list_del() can be called twice on
1495 req->rq_history_list, causing a kernel oops.
1499 Description: kptllnd_peer_check_sends()) ASSERTION(!in_interrupt()) failed
1500 Details : fix stack overflow in the distributed lock manager by defering
1501 export eviction after a failed AST to the elt thread instead of
1502 handling it in the dlm interpret routine.
1506 Description: Convert some comments to new format.
1507 Details : Update documenting comments to match doxygen conventions.
1511 Description: Grammar fixes.
1512 Details : A couple of trivial spelling fixes.
1516 Description: OSD_COUNTERS-mandatory
1517 Details : Make previously optional ->oti_{w,r}_locks sanity checks mandatory
1518 to simplify the code.
1522 Description: simplify cmm_device freeing logic.
1523 Details : Call cmm_device_free() in the failure path of cmm_device_alloc().
1527 Description: Add lockdep support to dt_object_operations locking interface.
1528 Details : Augment ->do_{read,write}_lock() prototypes with a `role'
1529 parameter indicating lock ordering. Update mdd code to use new
1534 Description: Introduce failloc constants for lockless IO tests.
1535 Details : Add two new failloc constants to test lockless IO. Only one of
1536 them in implemented---another is checked in yet to be landed
1541 Description: Add lockdep support for inode mutex.
1542 Details : Introduce and use new LOCK_INODE_MUTEX_PARENT() macro to be used
1543 in the situations where Lustre has to lock more than one inode
1548 Description: Add optional invariants checking support.
1549 Details : Add new LINVRNT() macro, optional on new --enable-invariants
1550 configure switch. This macro is to be used for consistency and
1551 sanity checks that are too expensive to be left in `production'
1556 Description: Zap lock->l_granted_mode with explicit LCK_MINMODE.
1557 Details : Use LCK_MINMODE rather than 0 to reset lock->l_granted_mode to
1562 Description: Add lockdep support for ldlm_lock and ldlm_resource.
1563 Details : Use spin_lock_nested() in (the only) situation where more than
1564 one ldlm_lock is locked simultaneously. Also, fix possible
1565 dead-lock in ldlm_lock_change_resource() by enforcing particular
1570 Description: Use struct ldlm_callback_suite in ldlm_lock_create().
1571 Details : Instead of specifying each ldlm_lock call-back through separate
1572 parameter, wrap them into struct ldlm_callback_suite.
1576 Description: Kill join_lru obd method and its callers.
1577 Details : CLIO uses lock weighting policy to keep locks over mmapped regions
1578 in memory---a requirement implemented through ->o_join_lru() obd
1579 method in HEAD. Remove this method and its users.
1583 Description: Add asynchronous ldlm ENQUEUE completion handler.
1584 Details : CLIO posts ENQUEUE requests asynchronously through ptlrpcd---a
1585 case that stock ldlm_completion_ast() cannot handle as it waits
1586 until lock is granted. Introduce new ldlm_completion_ast_async()
1587 for this. Also comment ldlm_completion_ast().
1591 Description: ldlm_error <-> errno conversion.
1592 Details : Add functions to map (rather arbitrary) between LDLM error codes
1593 and standard errno values. CLIO needs this to prevent LDLM specific
1594 constants from escaping ldlm and osc.
1598 Description: Kill unused ldlm_handle2lock_ns() function.
1599 Details : Kill unused ldlm_handle2lock_ns() function.
1603 Description: Add lu_ref support to ldlm_lock
1604 Details : lu_ref support for ldlm_lock and ldlm_resource. See lu_ref patch.
1605 lu_ref fields ->l_reference and ->lr_reference are added to ldlm_lock
1606 and ldlm_resource. LDLM interface has to be changed, because code that
1607 releases a reference on a lock, has to "know" what reference this is.
1608 In the most frequent case
1610 lock = ldlm_handle2lock(handle);
1612 LDLM_LOCK_PUT(lock);
1614 no changes are required. When any other reference (received _not_ from
1615 ldlm_handle2lock()) is released, LDLM_LOCK_RELEASE() has to be called
1616 instead of LDLM_LOCK_PUT().
1618 Arguably, changes are pervasive, and interface requires some discipline
1619 for proper use. On the other hand, it was very instrumental in finding
1620 a few leaked lock references.
1624 Description: Add ldlm_lock_addref_try().
1625 Details : Introduce ldlm_lock_addref_try() function (used by CLIO) that
1626 attempts to addref a lock that might be being canceled
1631 Description: Add ldlm_weigh_callback().
1632 Details : Add new ->l_weigh_ast() call-back to ldlm_lock. It is called
1633 by ldlm_cancel_shrink_policy() to estimate lock "value", instead of
1634 hard-coded `number of pages' logic.
1638 Description: Add lockdep annotations to llog code.
1639 Details : Use appropriately tagged _nested() locking calls in the places
1640 where llog takes more than one ->lgh_lock lock.
1644 Description: Add loi_kms_set().
1645 Details : Wrap kms updates into a helper function.
1649 Description: Constify instances of struct lsm_operations.
1650 Details : Constify instances of struct lsm_operations.
1654 Description: lu_conf support.
1655 Details : On a server, a file system object is uniquely identified
1656 by a fid, which is sufficient to locate and load all object
1657 state (inode). On a client, on the other hand, more data are
1658 necessary instantiate an object. Change lu_object_find() and
1659 friends to take additional `lu_conf' argument describing object.
1660 Typically this includes layout information.
1664 Description: lu_context fixes.
1665 Details : Introduce new lu_context functions that are needed on the client
1666 side, where some system threads (ptlrpcd) are shared by multiple
1667 modules, and so cannot be stopped during module shutdown.
1671 Description: Add start and stop methods to lu_device_type_operations.
1672 Details : Introduce two new methods in lu_device_type_operations, that are
1673 invoked when first instance of a given type is created and last one
1674 is destroyed respectively. This is need by CLIO.
1678 Description: Add lu_ref support to struct lu_device.
1679 Details : Add lu_ref support to lu_object and lu_device. lu_ref is used to
1680 track leaked references.
1684 Description: Introduce lu_kmem_descr.
1685 Details : lu_kmem_descr and its companion interface allow to create
1686 and destroy a number of kmem caches at once.
1690 Description: Fix lu_object finalization race.
1691 Details : Fix a race between lu_object_find() finding an object and its
1692 concurrent finalization. This race is (most likely) not possible
1693 on the server, but might happen on the client.
1697 Description: Introduce lu_ref interface.
1698 Details : lu_ref is a debugging module allowing to track references to
1699 a given object. It is quite cpu expensive, and has to be
1700 explicitly enabled with --enable-lu_ref. See usage description
1705 Description: Factor lu_site procfs stats into a separate function.
1706 Details : Separate lu_site stats printing code into a separate function
1707 that can be reused on a client.
1711 Description: Constify instances of struct {lu,dt,md}_device_operations.
1712 Details : Constify instances of struct {lu,dt,md}_device_operations.
1716 Description: Introduce struct md_site and move meta-data specific parts of
1717 struct lu_site here.
1718 Details : Move md-specific fields out of struct lu_site into special struct
1719 md_site, so that lu_site can be used on a client.
1723 Description: Kill mdd_lov_destroy().
1724 Details : Remove unused mdd code.
1728 Description: Add st_block checking to multistat.c.
1729 Details : Add st_block checking to multistat.c.
1733 Description: Add lu_ref support to struct obd_device.
1734 Details : Add lu_ref tracking to obd_device.
1738 Description: Kill obd_set_fail_loc().
1739 Details : Remove unused code.
1743 Description: Add special type for ptlrpc_request interpret functions.
1744 Details : Add lu_env parameter to ->rq_interpreter call-back. NULL is passed
1745 there. Actual usage will be in CLIO.
1749 Description: Replace RW_LOCK_UNLOCKED() macro with rwlock_init().
1750 Details : Replace RW_LOCK_UNLOCKED() with rwlock_init() as the former
1751 doesn't work with lockdep.
1755 Description: Add rwv.c test program.
1756 Details : New testing program exercising readv(2) and writev(2) (Qian).
1760 Description: Add sendfile.c test program.
1761 Details : New testing program exercising sendfile(2) (Jay).
1765 Description: Ratelimit a message that can be very frequent.
1766 Details : Ratelimit a memory allocation failure message that can
1771 Description: Use cdebug_show() in CDEBUG-style macros defined outside of libcfs.
1772 Details : Use cdebug_show() in CDEBUG-style macros defined outside of libcfs.
1776 Description: Liblustre build fixes.
1777 Details : Liblustre build fixes.
1781 Description: libcfs: add cfs_{need,cond}_resched() interface.
1782 Details : libcfs: add cfs_{need,cond}_resched() definition and
1783 implementations for Linux, NT, and liblustre.
1785 Severity : enhancement
1787 Description: More exported tunables for mballoc
1788 Details : Add support for tunable preallocation window and new tunables for
1789 large/small requests
1793 Description: Detect corruption of block bitmap and checking for preallocations
1794 Details : Checks validity of on-disk block bitmap. Also it does better
1795 checking of number of applied preallocations. When corruption is
1796 found, it turns filesystem readonly to prevent further corruptions.
1800 Description: (rw.c:1323:ll_read_ahead_pages()) ASSERTION(page_idx > ria->ria_stoff) failed
1801 Details : Once the unmatched stride IO mode is detected, shrink the stride-ahead
1802 window to 0. If it does hit cache miss, and read-pattern is still
1803 stride-io mode, does not reset the stride window, but also does not
1804 increase the stride window length in this case.
1808 Frequency : only for big-endian servers
1809 Description: Check if system is big-endian while mounting fs with extents feature
1810 Details : Mounting a filesystem with extents feature will fail on big-endian
1811 systems since ext3-based ldiskfs is not supported on big-endian
1812 systems. This can be over-riden with "bigendian_extents" mount option.
1814 --------------------------------------------------------------------------------
1816 2007-08-10 Cluster File Systems, Inc. <info@clusterfs.com>
1818 * Support for kernels:
1819 2.6.5-7.283 (SLES 9),
1820 2.6.9-55.EL (RHEL 4),
1821 2.6.16.46-0.14 (SLES 10),
1822 2.6.18.8 vanilla (kernel.org)
1823 * Client support for unpatched kernels:
1824 (see http://wiki.lustre.org/index.php?title=Patchless_Client)
1825 2.6.16 - 2.6.22 vanilla (kernel.org)
1826 * Due to recently discovered recovery problems, we do not recommend
1827 using patchless RHEL 4 clients with this or any earlier release.
1828 * Recommended e2fsprogs version: 1.39.cfs8
1829 * Note that reiserfs quotas are disabled on SLES 10 in this kernel.
1830 * Starting with this release, the ldiskfs backing filesystem required
1831 by Lustre is now in its own package, lustre-ldiskfs. This package
1832 should be installed. It is versioned separately from Lustre and
1833 may be released separately in future.
1837 Description: block reactivating mgc import until all deactivates complete
1838 Details : Fix race when failing back MDT/MGS to itself (testing)
1841 Frequency : at statup only
1843 Description: mds_lov_synchronize race leads to various problems
1844 Details : simultaneous MDT->OST connections at startup can cause the
1845 sync to abort, leaving the OSC in a bad state.
1847 Severity : enhancement
1849 Description: add optional extra BUILD_VERSION info
1850 Details : add a new environment variable (namely LUSTRE_VERS) which allows
1851 to override the lustre version.
1854 Frequency : 2.6.18 servers only
1856 Description: ll_kern_mount() doesn't release the module reference
1857 Details : The ldiskfs module reference count never drops down to 0
1858 because ll_kern_mount() doesn't release the module reference.
1863 Description: server LBUG when using old ost_num_threads parameter
1864 Details : Accept the old ost_num_threads parameter but warn that it
1865 is deprecated, and fix an off-by-one error that caused an LBUG.
1870 Description: Transient SCSI error results in persistent IO issue
1871 Details : iobuf->dr_error is not reinitialized to 0 between two
1875 Frequency : sometimes when underlying device returns I/O errors
1877 Description: OSTs not going read-only during write failures
1878 Details : OSTs are not remounted read-only when the journal commit threads
1879 get I/O errors because fsfilt_ext3 calls journal_start/stop()
1880 instead of the ext3 wrappers.
1884 Description: poor connect scaling with increasing client count
1885 Details : Don't run filter_grant_sanity_check for more than 100 exports
1886 to improve scaling for large numbers of clients.
1889 Frequency : SLES10 only
1891 Description: sanity-quota.sh quotacheck failed: rc = -22
1892 Details : Quotas cannot be enabled on SLES10.
1895 Frequency : liblustre clients only
1897 Description: getdirentries does not give error when run on compute nodes
1898 Details : getdirentries does not fail when the size specified as an argument
1899 is too small to contain at least one entry
1901 Severity : enhancement
1903 Description: Add LNET router traceability for debug purposes
1904 Details : If a checksum failure occurs with a router as part of the
1905 IO path, the NID of the last router that forwarded the bulk data
1906 is printed so it can be identified.
1911 Description: OST "spontaneously" evicts client; client has imp_pingable == 0
1912 Details : Due to a race condition, liblustre clients were occasionally
1913 evicted incorrectly.
1915 Severity : enhancement
1917 Description: lfs setstripe use optional parameters instead of postional
1920 Severity : enhancement
1922 Description: Nanosecond timestamp support for ldiskfs
1923 Details : The on-disk ldiskfs filesystem has added support for nanosecond
1924 resolution timestamps. There is not yet support for this at
1925 the Lustre filesystem level.
1928 Frequency : during server recovery
1930 Description: MDS failing to send precreate requests due to OSCC_FLAG_RECOVERING
1931 Details : request with rq_no_resend flag not awake l_wait_event if they get
1935 Frequency : nfs export on patchless client
1937 Description: connectathon hang when test nfs export over patchless client
1938 Details : Disconnected dentry cannot be found with lookup, so we do not need
1939 to unhash it or make it invalid
1942 Description: fix llapi_lov_get_uuids() to allow many OSTs to be returned
1943 Details: : Change llapi_lov_get_uuids() to read the UUIDs from /proc instead
1944 of using an ioctl. This allows lfsck for > 160 OSTs to succeed.
1949 Description: open req refcounting wrong on reconnect
1950 Details : If reconnect happened between getting open reply from server and
1951 call to mdc_set_replay_data in ll_file_open, we will schedule
1952 replay for unreferenced request that we are about to free.
1953 Subsequent close will crash in variety of ways.
1954 Check that request is still eligible for replay in
1955 mdc_set_replay_data().
1960 Description: disable writes to filesystem when reading health_check file
1961 Details : the default for reading the health_check proc file has changed
1962 to NOT do a journal transaction and write to disk, because this
1963 can cause reads of the /proc file to hang and block HA state
1964 checking on a healthy but otherwise heavily loaded system. It
1965 is possible to return to the previous behaviour during configure
1966 with --enable-health-write.
1968 Severity : enhancement
1970 Description: 64-bit inode version
1971 Details: : Add a on-disk 64-bit inode version for ext3 to track changes made
1972 to the inode. This will be required for version-based recovery.
1977 Description: MDS fails to start if a duplicate client export is detected
1978 Details : in some rare cases it was possible for a client to connect to
1979 an MDS multiple times. Upon recovery the MDS would detect this
1980 and fail during startup. Handle this more gracefully.
1982 Severity : enhancement
1984 Description: Add -o localflock option to simulate old noflock
1986 Details : This will achieve local-only flock/fcntl locks
1992 Description: log_commit_thread vs filter_destroy race leads to crash
1993 Details : Take import reference before releasing llog record semaphore
1998 Description: Wrong request locking in request set processing
1999 Details : ptlrpc_check_set wrongly uses req->rq_lock for proctect add to
2000 imp_delayed_list, in this place should be used imp_lock.
2003 Frequency : when reconnection
2005 Description: Grant Leak when osc reconnect to OST
2006 Details : When osc reconnect ost, OST(filter) should check whether it
2007 should grant more space to client by comparing fed_grant and
2008 cl_avail_grant, and return the granted space to client instead
2009 of "new granted" space, because client will call osc_init_grant
2010 to update the client grant space info.
2013 Frequency : when client reconnect to OST
2015 Description: Grant Leak when osc do resend and replay bulk write
2016 Details : When osc reconnect to OST, OST(filter)should clear grant info of
2017 bulk write request, because the grant info while be sync between
2018 OSC and OST when reconnect, and we should ignore the grant info
2019 these of resend/replay write req.
2024 Description: Grant space more than avaiable left space sometimes.
2025 Details : When then OST is about to be full, if two bulk writing from
2026 different clients came to OST. Accord the avaliable space of the
2027 OST, the first req should be permitted, and the second one
2028 should be denied by ENOSPC. But if the seconde arrived before
2029 the first one is commited. The OST might wrongly permit second
2030 writing, which will cause grant space > avaiable space.
2033 Frequency : when client is evicted
2035 Description: Grant might be wrongly erased when osc is evicted by OST
2036 Details : when the import is evicted by server, it will fork another
2037 thread ptlrpc_invalidate_import_thread to invalidate the
2038 import, where the grant will be set to 0. While the original
2039 thread will update the grant it got when connecting. So if
2040 the former happened latter, the grant will be wrongly errased
2041 because of this race.
2046 Description: Checking Stale with correct fid
2047 Details : ll_revalidate_it should uses de_inode instead of op_data.fid2
2048 to check whether it is stale, because sometimes, we want the
2049 enqueue happened anyway, and op_data.fid2 will not be initialized.
2051 Severity : enhancement
2053 Description: update patchless client
2054 Details : Add support for patchless client with 2.6.20, 2.6.21 and RHEL 5
2057 Frequency : only with 2.4 kernel
2059 Description: random memory corruption
2060 Details : size of struct ll_inode_info is to big for union inode.u and this
2061 can be cause of random memory corruption.
2066 Description: Memory leak in recovery
2067 Details : Lov_mds_md was not free in an error handler in mds_create_object.
2068 It should also check obd_fail before fsfilt_start, otherwise if
2069 fsfilt_start return -EROFS,(failover mds during mds recovery).
2070 then the req will return with repmsg->transno = 0 and rc = EROFS.
2071 and we met hit the assert LASSERT(req->rq_reqmsg->transno ==
2072 req->rq_repmsg->transno) in ptlrpc_replay_interpret. Fcc should
2073 be freed no matter whether fsfilt_commit success or not.
2076 Frequency : only with huge count clients
2078 Description: Prevents from taking the superblock lock in llap_from_page for
2080 Details : using LL_ORIGIN_REMOVEPAGE origin flag instead of LL_ORIGIN_UNKNOW
2081 for llap_from_page call in ll_removepage prevents from taking the
2082 superblock lock for a soon died page.
2087 Description: Not check open intent error before release open handle
2088 Details : in some rare cases, the open intent error is not checked before
2089 release open handle, which may cause
2090 ASSERTION(open_req->rq_transno != 0), because it tries to release
2091 the failed open handle.
2096 Description: Set cat log bitmap only after create log success.
2097 Details : in some rare cases, the cat log bitmap is set too early. and it
2098 should be set only after create log success.
2102 Description: Accessing a block bevice can re-enable I/O when Lustre is
2103 tearing down a device.
2104 Details : dev_clear_rdonly(bdev) must be called in kill_bdev() instead of
2109 Description: service threads may hog cpus when there are a lot of requests
2110 Details : Insert cond_resched to give other threads a chance to use some CPU
2115 Description: the cat log was not initialized in recovery
2116 Details : When mds(mgs) do recovery, the tgt_count might be zero, so the
2117 unlink log on mds will not be initialized until mds post
2118 recovery. And also in mds post recovery, the unlink log will
2119 initialization will be done asynchronausly, so there will be race
2120 between add unlink log and unlink log initialization.
2124 Description: brw_stats were being printed incorrectly
2125 Details : brw_stats were being printed as log2 but all of them were not
2126 recorded as log2. Also remove some code duplication arising from
2127 filter_tally_{read,write}.
2131 Frequency : rare, only in recovery.
2132 Description: ASSERTION(req->rq_type != LI_POISON) failed
2133 Details : imp_lock should be held while iterating over imp_sending_list for
2134 prevent destroy request after get timeout in ptlrpc_queue_wait.
2138 Description: replay-single.sh test 52 fails
2139 Details : A lock's skiplist need to be cleanup when it being unlinked
2140 from its resource list.
2144 Description: Short directio read returns full requested size rather than
2146 Details : Direct I/O operations should return actual amount of bytes
2147 transferred rather than requested size.
2149 Severity : enhancement
2151 Description: metadata RPC reduction (e.g. for rm performance)
2152 Details : decrease the amount of synchronous RPC between clients and servers
2153 by canceling conflicing lock before the operation on the client
2154 and packing thier handles into the main operation RPC to server.
2156 Severity : enhancement
2158 Description: Async OSC create to avoid the blocking unnecessarily.
2159 Details : If a OST has no remain object, system will block on the creating
2160 when need to create a new object on this OST. Now, ways use
2161 pre-created objects when available, instead of blocking on an
2162 empty osc while others are not empty. If we must block, we block
2163 for the shortest possible period of time.
2167 Description: improve handling recoverable errors
2168 Details : if request processig with error which can be recoverable on server
2169 request should be resend, otherwise page released from cache and
2172 Severity : enhancement
2174 Description: refine locking for avoid write wrong info into lov_objid file
2175 Details : fix possible races with add new target and write/update data in
2178 --------------------------------------------------------------------------------
2180 2007-05-03 Cluster File Systems, Inc. <info@clusterfs.com>
2185 Frequency : on some architectures
2187 Description: 1.6 client sometimes fails to mount from a 1.4 MDT
2188 Details : Uninitialized flags sometimes cause configuration commands to
2192 Frequency : patchless clients only
2194 Description: missing __iget() symbol export
2195 Details : The __iget() symbol export is missing. To avoid the need for
2196 this on patchless clients the deathrow inode reaper is turned
2197 off, and we depend on the VM to clean up old inodes. This
2198 dependency was during via the fix for bug 12181.
2200 --------------------------------------------------------------------------------
2202 2007-04-19 Cluster File Systems, Inc. <info@clusterfs.com>
2204 * CONFIGURATION CHANGE. This version of Lustre WILL NOT
2205 INTEROPERATE with older versions automatically. In many cases a
2206 special upgrade step is needed. Please read the
2207 user documentation before upgrading any part of a live system.
2208 * WIRE PROTOCOL CHANGE from previous 1.6 beta versions. This
2209 version will not interoperate with 1.6 betas before beta5 (1.5.95).
2210 * WARNING: Lustre configuration and startup changes are required with
2211 this release. See https://mail.clusterfs.com/wikis/lustre/MountConf
2215 Severity : enhancement
2217 Description: MountConf
2218 Details : Lustre configuration is now managed via mkfs and mount
2219 commands instead of lmc and lconf. New obd types (MGS, MGC)
2220 are added for dynamic configuration management. See
2221 https://mail.clusterfs.com/wikis/lustre/MountConf for
2224 Severity : enhancement
2226 Description: dynamic OST addition
2227 Details : OSTs can now be added to a live filesystem
2229 Severity : enhancement
2231 Description: startup order invariance
2232 Details : MDTs and OSTs can be started in any order. Clients only
2233 require the MDT to complete startup.
2235 Severity : enhancement
2237 Description: parallel, asynchronous orphan cleanup
2238 Details : orphan cleanup is now performed in separate threads for each
2239 OST, allowing parallel non-blocking operation.
2241 Severity : enhancement
2243 Description: optimized stripe assignment
2244 Details : stripe assignments are now made based on ost space available,
2245 ost previous usage, and OSS previous usage, in order to try
2246 to optimize storage space and networking resources.
2248 Severity : enhancement
2250 Description: Permanently set tunables
2251 Details : All writable /proc/fs/lustre tunables can now be permanently
2252 set on a per-server basis, at mkfs time or on a live system.
2254 Severity : enhancement
2256 Description: Lustre message v2
2257 Details : Add lustre message format v2.
2259 Severity : enhancement
2261 Description: client OST exclusion list
2262 Details : Clients can be started with a list of OSTs that should be
2263 declared "inactive" for known non-responsive OSTs.
2267 Description: SPEC SFS validation failure on NFS v2 over lustre.
2268 Details : Changes the blocksize for regular files to be 2x RPC size,
2269 and not depend on stripe size.
2271 Severity : enhancement
2273 Description: Multiple MD RPCs in flight.
2274 Details : Further unserialise some read-only MDS RPCs - learn about intents.
2275 To avoid overly-overloading MDS, introduce a limit on number of
2276 MDS RPCs in flight for a single client and add /proc controls
2277 to adjust this limit.
2279 Severity : enhancement
2281 Description: client read/write statistics
2282 Details : Add client read/write call usage stats for performance
2283 analysis of user processes.
2284 /proc/fs/lustre/llite/*/offset_stats shows non-sequential
2285 file access. extents_stats shows chunk size distribution.
2286 extents_stats_per_process show chunk size distribution per
2289 Severity : enhancement
2291 Description: mds statistics
2292 Details : Add detailed mds operations statistics in
2293 /proc/fs/lustre/mds/*/stats.
2297 Description: Failure of copying files with lustre special EAs.
2298 Details : Client side always return success for setxattr call for lustre
2299 special xattr (currently only "trusted.lov").
2304 Description: Refcount LNET uuids
2305 Details : The global LNET uuid list grew linearly with every startup;
2306 refcount repeated list entries instead of always adding to
2309 Severity : enhancement
2311 Description: Dynamic service threads
2312 Details : Within a small range, start extra service threads
2313 automatically when the request queue builds up.
2316 Frequency : mixed-endian client/server environments
2318 Description: mixed-endian crashes
2319 Details : The new msg_v2 system had some failures in mixed-endian
2322 Severity : enhancement
2324 Description: Easy OST removal
2325 Details : OSTs can be permanently deactivated with e.g. 'lctl
2326 conf_param lustre-OST0001.osc.active=0'
2328 Severity : enhancement
2330 Description: MGS proc entries
2331 Details : Added basic proc entries for the MGS showing what filesystems
2334 Severity : enhancement
2336 Description: provide MGS failover
2337 Details : Added config lock reacquisition after MGS server failover.
2339 Severity : enhancement
2341 Description: add Linux 2.4 support
2342 Details : Added support for RHEL 2.4.21 kernel for 1.6 servers and clients
2346 Description: a large application tries to do I/O to the same resource and dies
2347 in the middle of it.
2348 Details : Check the req->rq_arrival time after the call to
2349 ost_brw_lock_get(), but before we do anything about
2350 processing it & sending the BULK transfer request. This
2351 should help move old stale pending locks off the queue as
2352 quickly as obd_timeout.
2355 Frequency : when an incorrect nid is specified during startup
2357 Description: ptlrpc connect to non-existant node causes kernel crash
2358 Details : LNET can't be re-entered from an event callback, which
2359 happened when we expire a message after the export has been
2360 cleaned up. Instead, hand the zombie cleanup off to another
2363 Severity : enhancement
2365 Description: plain/inodebits lock performance improvement
2366 Details : Grouping plain/inodebits in granted list by their request modes
2367 and bits policy, thus improving the performance of search through
2371 Frequency : only if OST filesystem is corrupted
2373 Description: client incorrectly hits assertion in ptlrpc_replay_req()
2374 Details : for a short time RPCs with bulk IO are in the replay list,
2375 but replay of bulk IOs is unimplemented. If the OST filesystem
2376 is corrupted due to disk cache incoherency and then replay is
2377 started it is possible to trip an assertion. Avoid putting
2378 committed RPCs into the replay list at all to avoid this issue.
2381 Frequency : liblustre (e.g. catamount) on a large cluster with >= 8 OSTs/OSS
2383 Description: System hang on startup
2384 Details : This bug allowed the liblustre (e.g. catamount) client to
2385 return to the app before handling all startup RPCs. This
2386 could leave the node unresponsive to lustre network traffic
2387 and manifested as a server ptllnd timeout.
2389 Severity : enhancement
2391 Description: Add "/proc/sys/lustre/debug_peer_on_timeout"
2392 Details : liblustre envirable: LIBLUSTRE_DEBUG_PEER_ON_TIMEOUT
2393 boolean to control whether to print peer debug info when a
2394 client's RPC times out.
2397 Frequency : only for kernels with patches from Lustre below 1.4.3
2399 Description: Remove old rdonly API
2400 Details : Remove old rdonly API which unused from at least lustre 1.4.3
2403 Frequency : only for devices with external journals
2405 Description: Set external device read-only also
2406 Details : During a commanded failover stop, we set the disk device
2407 read-only while the server shuts down. We now also set any
2408 external journal device read-only at the same time.
2411 Frequency : when upgrading from 1.4 while trying to change parameters
2413 Description: The wrong (new) MDC name was used when setting parameters for
2414 upgraded MDT's. Also allows changing of OSC (and MDC)
2415 parameters if --writeconf is specified at tunefs upgrade time.
2418 Frequency : when setting specific ost indicies
2420 Description: QOS code breaks on skipped indicies
2421 Details : Add checks for missing OST indicies in the QOS code, so OSTs
2422 created with --index need not be sequential.
2424 Severity : enhancement
2426 Description: Add uninit_groups feature to ldiskfs2 to speed up e2fsck
2427 Details : The uninit_groups feature works in conjunction with the kernel
2428 filesystem code (ldiskfs2 only) and e2fsprogs-1.39-cfs6 to speed
2429 up the pass1 processing of e2fsck. This is a read-only feature
2430 in ldiskfs2 only, so older kernels and current ldiskfs cannot
2431 mount filesystems that have had this feature enabled.
2433 Severity : enhancement
2435 Description: Improve multi-block allocation algorithm to avoid fragmentation
2436 Details : The mballoc3 code (ldiskfs2 only) adds new mechanisms to improve
2437 allocation locality and avoid filesystem fragmentation.
2439 ------------------------------------------------------------------------------
2441 2007-04-01 Cluster File Systems, Inc. <info@clusterfs.com>
2443 * Support for kernels:
2444 2.4.21-47.0.1.EL (RHEL 3)
2445 2.6.5-7.283 (SLES 9)
2446 2.6.9-42.0.10.EL (RHEL 4)
2447 2.6.12.6 vanilla (kernel.org)
2448 2.6.16.27-0.9 (SLES 10)
2449 * Recommended e2fsprogs version: 1.39.cfs5
2451 * Note that reiserfs quotas are disabled on SLES 10 in this kernel
2455 Frequency : occasional, depends on client load and configuration
2456 Bugzilla : 12181, 12203
2457 Description: data loss for recently-modified files
2459 Details : In some cases it is possible that recently written or created
2460 files may not be written to disk in a timely manner (this should
2461 normally be within 30s unless client IO load is very high).
2462 The problem appears as zero-length files or files that are a
2463 multiple of 1MB in size after a client crash or client eviction
2464 that are missing data at the end of the file.
2466 This problem is more likely to be hit on clients where files are
2467 repeatedly created and unlinked in the same directory, clients
2468 have a large amount of RAM, have many CPUs, the filesystem has
2469 many OSTs, the clients are rebooted frequently, and/or the files
2470 are not accessed by other nodes after being written.
2472 The presence of the problem can be detected by looking at
2473 /proc/sys/fs/inode-state. If the first number (nr_inodes) is
2474 smaller than the second (nr_unused) then dirty files will not
2475 be flushed automatically to disk. "sync; sleep 10" should be
2476 run several times on the node before unmounting it to update
2477 Lustre (this is also safe to run on nodes without this problem).
2479 There is also a related kernel bug in the RHEL4 4 2.6.9 kernel
2480 that can cause this same problem, so customers using that kernel
2481 also need to update the kernel in addition to Lustre. In order
2482 to properly fix this bug, the RHEL3 2.4.21 kernel is also updated.
2484 It is normal that files written just before a client crash (less
2485 than 30s) may not yet have been flushed to disk, even for local
2489 Frequency : frequent on thin XT3 nodes
2491 Description: UUID collision on thin XT3 Linux nodes
2492 Details : UUIDs on Compute Node Linux XT3 nodes were not generated
2493 randomly, since we relied on an insufficiently-seeded PRNG.
2498 Description: OSS hangs after "All ost request buffers busy"
2499 Details : A deadlock between quota and journal operations caused OSS
2500 hangs after printing "All ost request buffers busy."
2503 Frequency : always on liblustre builds
2505 Description: Cleanup compiler warnings on liblustre
2508 Frequency : always on liblustre builds on XT3
2510 Description: LC_CONFIG_CDEBUG don't run while build liblustre on XT3.
2514 Description: Addition of EXT3_FEATURE_RO_COMPAT_DIR_NLINKS flag for
2515 > 32000 subdirectories
2516 Details : Add EXT3_FEATURE_RO_COMPAT_DIR_NLINK flag to
2517 EXT3_FEATURE_RO_COMPAT_SUPP. This flag will be set whenever
2518 subdirectory count crosses 32000. This will aid e2fsck to
2519 correctly handle more than 32000 subdirectories.
2522 Frequency : liblustre (e.g. catamount) on a large cluster with >= 8 OSTs/OSS
2524 Description: System hang on startup
2525 Details : This bug allowed the liblustre (e.g. catamount) client to
2526 return to the app before handling all startup RPCs. This
2527 could leave the node unresponsive to lustre network traffic
2528 and manifested as a server ptllnd timeout.
2530 Severity : enhancement
2532 Description: Add "/proc/sys/lustre/debug_peer_on_timeout"
2533 (liblustre envirable: LIBLUSTRE_DEBUG_PEER_ON_TIMEOUT)
2534 boolean to control whether to print peer debug info when a
2535 client's RPC times out.
2540 Description: make O_SYNC working on 2.6 kernels
2541 Details : 2.6 kernels use different method for mark pages for write,
2542 so need add a code to lustre for O_SYNC work.
2547 Description: Failure to close file and release space on NFS
2548 Details : Put inode details into lock acquired in ll_intent_file_open.
2549 Use mdc_intent_lock in ll_intent_open to properly
2550 detect all kind of errors unhandled by mdc_enqueue.
2555 Description: proc file read during shutdown sometimes raced obd removal,
2557 Details : Add lock to prevent obd access after proc file removal.
2560 Frequency : Only for files larger than 4GB on 32-bit clients.
2562 Description: improperly doing page alignment of locks
2563 Details : Modify lustre core code to use CFS_PAGE_* defines instead of
2564 PAGE_*. Make CFS_PAGE_MASK a 64-bit mask.
2569 Description: RPCs being resent when they shouldn't be
2570 Details : Some RPCs that should not be resent are being resent. This
2571 can cause inconsistencies in the RPC state machine. Do not
2572 resend such requests.
2575 Frequency : rare, only with NFS export
2577 Description: Crash on NFS re-export node
2578 Details : under very unusual load conditions an assertion is hit in
2579 ll_intent_file_open()
2582 Frequency : only if OST filesystem is corrupted
2584 Description: client incorrectly hits assertion in ptlrpc_replay_req()
2585 Details : for a short time RPCs with bulk IO are in the replay list,
2586 but replay of bulk IOs is unimplemented. If the OST filesystem
2587 is corrupted due to disk cache incoherency and then replay is
2588 started it is possible to trip an assertion. Avoid putting
2589 committed RPCs into the replay list at all to avoid this issue.
2594 Description: large O_DIRECT requests fail under memory pressure/fragmentation
2595 Details : Large single O_DIRECT read and write calls can fail to allocate
2596 a sufficiently large buffer to process the request. In case of
2597 allocation failure the allocation is retried with a smaller
2598 buffer and broken into smaller requests.
2600 Severity : enhancement
2602 Description: Add -o localflock option to simulate old noflock behaviour.
2603 Details : This will achieve local-only flock/fcntl locks coherentness.
2608 Description: versioning check is incomplete
2609 Details : Checking the version difference of client vs. server, report
2610 error if the gap is too big.
2615 Description: add support PG_writeback bit
2616 Details : add support for PG_writeback bit for Lustre, for more carefull
2617 work with page cache in 2.6 kernel. This also fix some deadlocks
2618 and remove hack for work O_SYNC with 2.6 kernel.
2620 Severity : enhancement
2622 Description: Add uninit_groups feature to ldiskfs2 to speed up e2fsck
2623 Details : The uninit_groups feature works in conjunction with the kernel
2624 filesystem code (ldiskfs2 only) and e2fsprogs-1.39-cfs6 to speed
2625 up the pass1 processing of e2fsck. This is a read-only feature
2626 in ldiskfs2 only, so older kernels and current ldiskfs cannot
2627 mount filesystems that have had this feature enabled.
2629 Severity : enhancement
2631 Description: Improve multi-block allocation algorithm to avoid fragmentation
2632 Details : The mballoc3 code (ldiskfs2 only) adds new mechanisms to improve
2633 allocation locality and avoid filesystem fragmentation.
2635 ------------------------------------------------------------------------------
2637 2007-02-09 Cluster File Systems, Inc. <info@clusterfs.com>
2639 * Support for kernels:
2640 2.6.9-42.0.3.EL (RHEL 4)
2641 2.6.5-7.276 (SLES 9)
2642 2.4.21-47.0.1.EL (RHEL 3)
2643 2.6.12.6 vanilla (kernel.org)
2644 2.6.16.21-0.8 (SLES10)
2645 * Recommended e2fsprogs version: 1.39.cfs2-0
2647 * The backwards-compatible /proc/sys/portals symlink has been removed
2648 in this release. Before upgrading, please ensure that you change
2649 any configuration scripts or /etc/sysctl.conf files that access
2650 /proc/sys/portals/* or sysctl portals.* to use the corresponding
2651 entry in /proc/sys/lnet or sysctl lnet.*. This change can be made
2652 in advance of the upgrade on any system running Lustre 1.4.6 or
2653 newer, since /proc/sys/lnet was added in that version.
2654 * Note that reiserfs quotas are disabled on SLES 10 in this kernel
2658 Frequency : only when quota is used
2660 Description: avoid scanning export list for quota master
2661 Details : Change the algorithms to avoid scanning export list in order
2662 to improve the efficiency.
2665 Frequency : MDS failover only, very rarely
2667 Description: "went back in time" messages on mds failover
2668 Details : The greatest transno may be lost when the current operation
2669 finishes with an error (transno==0) and the client's last_rcvd
2670 record is over-written. Save the greatest transno in the
2671 mds_last_transno for this case.
2674 Frequency : always for specific kernels and striping counts
2676 Description: client may get "Matching packet too big" without ACL support
2677 Details : Clients compiled without CONFIG_FS_POSIX_ACL get an error message
2678 when trying to access files in certain configurations. The
2679 clients should in fact be denied when mounting because they do
2680 not understand ACLs.
2683 Frequency : Cray XT3 with more than 4000 clients and multiple jobs
2685 Description: many clients connecting with IO in progress causes connect timeouts
2686 Details : Avoid synchronous journal commits to avoid delays caused by many
2687 clients connecting/disconnecting when bulk IO is in progress.
2688 Queue liblustre connect requests on OST_REQUEST_PORTAL instead of
2689 OST_IO_PORTAL to avoid delays behind potentially many pending
2693 Frequency : occasionally with multiple writers to a single file
2695 Description: shared writes to file may result in wrong size reported by stat()
2696 Details : Allow growing of kms when extent lock is cancelled
2699 Frequency : always with random mmap IO to multi-striped file
2701 Description: mmap write might be lost if we are writing to a 'hole' in stripe
2702 Details : Only if the hole is at the end of OST object so that kms is too
2703 small. Fix is to increase kms accordingly in ll_nopage.
2706 Frequency : rare, only if OST filesystem is inconsistent with MDS filesystem
2708 Description: writes to a missing object would leak memory on the OST
2709 Details : If there is an inconsistency between the MDS and OST filesystems,
2710 such that the MDS references an object that doesn't exist, writes
2711 to that object will leak memory due to incorrect cleanup in the
2712 error handling path, eventually running out of memory on the OST.
2717 Description: Creating too long symlink causes lustre errors
2718 Details : Check symlink and name lengths before sending requests to MDS.
2721 Frequency : only if flock is enabled (not on by default)
2723 Description: posix locks not released on fd closure on 2.6.9+
2724 Details : We failed to add posix locks to list of inode locks on 2.6.9+
2725 kernels, this caused such locks not to be released on fd close and
2726 then assertions on fs unmount about still used locks.
2729 Frequency : MDS failover only, very rarely
2731 Description: clients may get ASSERTION(granted_lock != NULL)
2732 Details : When request was taking a long time, and a client was resending
2733 a getattr by name lock request. The were multiple lock requests
2734 with the same client lock handle and
2735 mds_getattr_name->fixup_handle_for_resent_request found one of the
2736 lock handles but later failed with ASSERTION(granted_lock != NULL).
2741 Description: handle->h_buffer_credits > 0, assertion failure
2742 Details : h_buffer_credits is zero after truncate, causing assertion
2743 failure. This patch extends the transaction or creates a new
2747 Frequency : NFS re-export or patchless client
2748 Bugzilla : 11179, 10796
2749 Description: Crash on NFS re-export node (__d_move)
2750 Details : We do not want to hash the dentry if we don't have a lock.
2751 But if this dentry is later used in d_move, we'd hit uninitialised
2752 list head d_hash, so we just do this to init d_hash field but
2753 leave dentry unhashed.
2756 Frequency : NFS re-export or patchless client
2758 Description: NFS exports has problem with symbolic link
2759 Details : lustre client didn't properly install dentry when re-exported
2760 to NFS or running patchless client.
2763 Frequency : NFS re-export or patchless client
2765 Description: Various nfs/patchless fixes.
2766 Details : fixes reuse disconected alias for lookup process - this fixes
2767 warning "find_exported_dentry: npd != pd",
2768 fix permission error with open files at nfs.
2769 fix apply umask when do revalidate.
2772 Frequency : occasional
2774 Description: Crash on NFS re-export node
2775 Details : calling clear_page() on the wrong pointer triggered oops in
2776 generic_mapping_read().
2779 Frequency : rarely, using O_DIRECT IO
2781 Description: unaligned directio crashes client with LASSERT
2782 Details : check for unaligned buffers before trying any requests.
2785 Frequency : rarely, using CFS RAID5 patches in non-standard kernel series
2787 Description: stale data returned from RAID cache
2788 Details : If only a small amount of IO is done to the RAID device before
2789 reading it again it is possible to get stale data from the RAID
2790 cache instead of reading it from disk.
2793 Frequency : always for sles10 kernel
2795 Description: sles10 support
2796 Details : ll_follow_link: compile fixes and using of nd_set_link
2797 under newer kernels.
2800 Frequency : depends on arch, kernel and compiler version, always on sles10
2803 Description: recursive or deep enough symlinks cause stack overflow
2804 Details : getting rid of large stack-allocated variable in
2808 Frequency : depends on hardware
2810 Description: lustre write performance loss in the SLES10 kernel
2811 Details : the performance loss is caused by using of write barriers in the
2812 ext3 code. The SLES10 kernel turns barrier support on by
2813 default. The fix is to undo that change for ldiskfs.
2815 ------------------------------------------------------------------------------
2817 2006-12-09 Cluster File Systems, Inc. <info@clusterfs.com>
2819 * Support for kernels:
2820 2.6.9-42.0.3EL (RHEL 4)
2821 2.6.5-7.276 (SLES 9)
2822 2.4.21-47.0.1.EL (RHEL 3)
2823 2.6.12.6 vanilla (kernel.org)
2827 Frequency : quota enabled and large files being deleted
2829 Description: releasing more than 4GB of quota at once hangs OST
2830 Details : If a user deletes more than 4GB of files on a single OST it
2831 will cause the OST to spin in an infinite loop. Release
2832 quota in < 4GB chunks, or use a 64-bit value for 1.4.7.1+.
2837 Description: statfs data retrieved from /proc may be stale or zero
2838 Details : When reading per-device statfs data from /proc, in the
2839 {kbytes,files}_{total,free,avail} files, it may appear
2840 as zero or be out of date.
2843 Frequency : systems with MD RAID1 external journal devices
2845 Description: lconf's call to blkid is confused by RAID1 journal devices
2846 Details : Use the "blkid -l" flag to locate the MD RAID device instead
2847 of returning all block devices that match the journal UUID.
2850 Frequency : always, for aggregate stripe size over 4GB
2852 Description: "lfs setstripe" fails assertion when setting 4GB+ stripe width
2853 Details : Using "lfs setstripe" to set stripe size * stripe count over 4GB
2854 will fail the kernel with "ASSERTION(lsm->lsm_xfersize != 0)"
2857 Frequency : always if "lfs find" used on a local file/directory
2859 Description: "lfs find" segfaults if used on a local file/directory
2860 Details : The case where a directory component was not specified wasn't
2861 handled correctly. Handle this properly.
2864 Frequency : always on ppc64
2866 Description: the write to an ext3 filesystem mounted with mballoc got stuck
2867 Details : ext3_mb_generate_buddy() uses find_next_bit() which does not
2868 perform endianness conversion.
2871 Frequency : rarely (truncate to non-zero file size after write under load)
2872 Bugzilla : 10730, 10687
2873 Description: Files padded with zeros to next 4K multiple
2874 Details : With filesystems mounted using the "extents" option (2.6 kernels)
2875 it is possible that files that are truncated to a non-zero size
2876 immediately after being written are filled with zero bytes beyond
2877 the truncated size. No file data is lost.
2879 Severity : enhancement
2881 Description: Allow recovery/failover for liblustre clients.
2882 Details : liblustre clients were unaware of failover configurations until
2885 Severity : enhancement
2887 Description: user file locks should fail when not mounting with flock option
2888 Details : Set up an error-returning stub in ll_file_operations.lock field
2889 to prevent incorrect behaviour when client is mounted without
2890 flock option. Also, set up properly f_op->flock field for
2894 Frequency : always on ia64
2896 Description: "lfs df" loops on printing out MDS statfs information
2897 Details : The obd_ioctl_data was not initialized and in some systems
2898 this caused a failure during the ioctl that did not return
2899 an error. Initialize the struct and return an error on failure.
2902 Frequency : SLES 9 only
2904 Description: Error of copying files with lustre special EAs as root
2905 Details : Client side always return success for setxattr call for lustre
2906 special xattr (currently only "trusted.lov").
2909 Frequency : rarely on clusters with both ia64+i386 clients
2911 Description: ia64+i686 clients doing shared IO on the same file may LBUG
2912 Details : In rare cases when both ia64+i686 (or other mixed-PAGE_SIZE)
2913 clients are doing concurrent writes to the same file it is
2914 possible that the ia64 clients may LASSERT because the OST
2915 extent locks are not PAGE_SIZE aligned. Ensure that grown
2916 locks are always aligned on the request boundary.
2919 Frequency : specific use, occasional
2921 Description: Overwriting in use executable truncates on-disk binary image
2922 Details : If one node attempts to overwrite an executable in use by
2923 another node, we now correctly return ETXTBSY instead of
2924 truncating the file.
2926 Severity : enhancement
2928 Description: Async OSC create to avoid the blocking unnecessarily.
2929 Details : If a OST has no remain object, system will block on the creating
2930 when need to create a new object on this OST. Now, ways use
2931 pre-created objects when available, instead of blocking on an
2932 empty osc while others are not empty. If we must block, we block
2933 for the shortest possible period of time.
2938 Description: chmod on Lustre root is propagated to other clients
2939 Details : Re-validate root's dentry in ll_lookup_it to avoid having it
2940 invalid by the follow_mount time.
2943 Frequency : liblustre clients only
2945 Description: Race in 'instant cancel' lock handling could lead to such locks
2946 never to be granted in case of SMP MDS
2947 Details : Do not destroy not yet granted but cbpending locks in
2951 Frequency : replay/resend of open
2953 Description: non null lock assetion failure in mds_intent_policy
2954 Details : Trying to replay/resend lockless open requests resulted in
2955 mds_open() returning 0 with no lock. Now it sets a flag if
2956 a lock is going to be returned.
2958 Severity : enhancement
2960 Description: Checksum enhancements
2961 Details : New checksum enhancements allow for resending RPCs that failed
2964 Severity : enhancement
2966 Description: Tunables on number of dirty pages in cacche
2967 Details : Allow to set limit on number of dirty pages cached.
2972 Description: client crash on unmount - lock still has references
2973 Details : In some error handling cases it was possible to leak a lock
2974 reference on a client while accessing a file. This was not
2975 harmful to the client during operation, but would cause the
2976 client to crash when the filesystem is unmounted.
2979 Frequency : specific case, rare
2981 Description: ETXTBSY on mds though file not in use by client
2982 Details : ETXTBSY is no longer incorrectly returned when attempting to
2983 chmod or chown a directory that the user previously tried to
2984 execute or a currently-executing binary.
2987 Frequency : extremely rare except on liblustre-based clients
2989 Description: Lustre space not freed when files are deleted
2990 Details : Clean up open-unlinked files after client eviction. Previously
2991 the unlink was skipped and the files remained as orphans.
2996 Description: OST failure "would be an LBUG" in waiting_locks_callback()
2997 Details : In some cases it was possible to send a blocking callback to
2998 a client doing a glimpse, even though that client didn't get
2999 a lock granted. When the glimpse lock is cancelled on the OST
3000 the freed lock is left on the waiting list and corrupted the list.
3003 Frequency : all core dumps
3005 Description: Broke core dumps to lustre
3006 Details : Negative dentry may be unhashed if parent does not have UPDATE
3007 lock, but some callers, e.g. do_coredump, expect dentry to be
3008 hashed after successful create, hash it in ll_create_it.
3010 ------------------------------------------------------------------------------
3012 2006-09-13 Cluster File Systems, Inc. <info@clusterfs.com>
3014 * Support for kernels:
3015 2.6.9-42.0.2.EL (RHEL 4)
3016 2.6.5-7.276 (SLES 9)
3017 2.4.21-40.EL (RHEL 3)
3018 2.6.12.6 vanilla (kernel.org)
3022 Frequency : always on RHEL 3
3024 Description: Number of open files grows over time
3025 Details : The number of open files grows over time, whether or not
3026 Lustre is started. This was due to a filp leak introduced
3027 by one of our kernel patches.
3029 ------------------------------------------------------------------------------
3031 08-20-2006 Cluster File Systems, Inc. <info@clusterfs.com>
3033 * Support for kernels:
3034 2.6.9-42.EL (RHEL 4)
3035 2.6.5-7.276 (SLES 9)
3036 2.4.21-40.EL (RHEL 3)
3037 2.6.12.6 vanilla (kernel.org)
3042 Bugzilla : 5719, 9635, 9792, 9684
3043 Description: OST (or MDS) trips assertions in (re)connection under heavy load
3044 Details : If a server is under heavy load and cannot reply to new
3045 connection requests before the client resends the (re)connect,
3046 the connection handling code can behave badly if two service
3047 threads are concurrently handing separate (re)connections from
3048 the same client. Add better locking to the connection handling
3049 code, and ensure that only a single connection will be processed
3050 for a given client UUID, even if the lock is dropped.
3052 Severity : enhancement
3054 Description: add TCP zero-copy support to kernel
3055 Details : Add support to the kernel TCP stack to allow zero-copy bulk
3056 sends if the hardware supports scatter-gather and checksumming.
3057 This allows socklnd to do client-write and server-read more
3058 efficiently and reduce CPU utilization from skbuf copying.
3061 Frequency : only if NFS exporting from client
3063 Description: NULL pointer deref in ll_iocontrol() if chattr mknod file
3064 Details : If setting attributes on a file created under NFS that had
3065 never been opened it would be possible to oops the client
3066 if the file had no objects.
3069 Frequency : always for liblustre
3071 Description: liblustre client does MDS+OSTs setattr RPC for each write
3072 Details : When doing a write from a liblustre client, the client
3073 incorrectly issued an RPC to the MDS and each OST the file was
3074 striped over in order to update the timestamps. When writing
3075 with small chunks and many clients this could overwhelm the MDS
3076 with RPCs. In all cases it would slow down the write because
3077 these RPCs are unnecessary.
3079 Severity : enhancement
3081 Description: allow number of MDS service threads to be changed at module load
3082 Details : It is now possible to change the number of MDS service threads
3083 running. Adding "options mds mds_num_threads={N}" to the MDS's
3084 /etc/modprobe.conf will set the number of threads for the next
3085 time Lustre is restarted (assuming the "mds" module is also
3086 reloaded at that time). The default number of threads will
3087 stay the same, 32 for most systems.
3092 Description: OST crash if filesystem is unformatted or corrupt
3093 Details : If an OST is started on a device that has never been formatted
3094 or if the filesystem is corrupt and cannot even mount then the
3095 error handling cleanup routines would dereference a NULL pointer.
3100 Description: NULL pointer deref in llap_from_page.
3101 Details : get_cache_page_nowait can return a page with NULL (or otherwise
3102 incorrect) mapping if the page was truncated/reclaimed while it was
3103 searched for. Check for this condition and skip such pages when
3104 doing readahead. Introduce extra check to llap_from_page() to
3105 verify page->mapping->host is non-NULL (so page is not anonymous).
3108 Frequency : Sometimes when using sys_sendfile
3110 Description: "page not covered by a lock" warnings from ll_readpage
3111 Details : sendfile called ll_readpage without right page locks present.
3112 Now we introduced ll_file_sendfile that does necessary locking
3113 around call to generic_file_sendfile() much like we do in
3117 Frequency : with certain MDS communication failures at client mount time
3119 Description: NULL pointer deref after failed client mount
3120 Details : a client connection request may delayed by the network layer
3121 and not be sent until after the PTLRPC layer has timed out the
3122 request. If the client fails the mount immediately it will try
3123 to clean up before the network times out the request. Add a
3124 reference from the request import to the obd device and delay
3125 the cleanup until the network drops the request.
3128 Frequency : occasionally during client (re)connect
3130 Description: assertion failure during client (re)connect
3131 Details : processing a client connection request may be delayed by the
3132 client or server longer than the client connect timeout. This
3133 causes the client to resend the connection request. If the
3134 original connection request is replied in this interval, the
3135 client may trip an assertion failure in ptlrpc_connect_interpret()
3136 which thought it would be the only running connect process.
3139 Frequency : only with obd_echo servers and clients that are rebooted
3141 Description: kernel BUG accessing uninitialized data structure
3142 Details : When running an obd_echo server it did not start the ping_evictor
3143 thread, and when a client was evicted an uninitialized data
3144 structure was accessed. Start the ping_evictor in the RPC
3145 service startup instead of the OBD startup.
3147 Severity : enhancement
3148 Bugzilla : 10193 (patchless)
3149 Description: Remove dependency on various unexported kernel interfaces.
3150 Details : No longer need reparent_to_init, exit_mm, exit_files,
3151 sock_getsockopt, filemap_populate, FMODE_EXEC, put_filp.
3154 Frequency : rare (only users of deprecated and unsupported LDAP config)
3156 Description: write_conf for zeroconf mount queried LDAP incorrectly for client
3157 Details : LDAP apparently contains 'lustreName' attributes instead of
3158 'name'. A simple remapping of the name is sufficient.
3161 Frequency : rare (only with non-default dump_on_timeout debug enabled)
3163 Description: waiting_locks_callback trips kernel BUG if client is evicted
3164 Details : Running with the dump_on_timeout debug flag turned on makes
3165 it possible that the waiting_locks_callback() can try to dump
3166 the Lustre kernel debug logs from an interrupt handler. Defer
3167 this log dumping to the expired_lock_main() thread.
3169 Severity : enhancement
3171 Description: Support NFS exporting on 2.6 kernels.
3172 Details : Implement non-rawops metadata methods for NFS server to use without
3173 changing NFS server code.
3176 Frequency : very rare (synthetic metadata workload only)
3178 Description: two racing renames might cause an MDS thread to deadlock
3179 Details : Running the "racer" program may cause one MDS thread to rename
3180 a file from being the source of a rename to being the target of
3181 a rename at exactly the same time that another thread is doing
3182 so, and the second thread has already enqueued these locks after
3183 doing a lookup of the target and is trying to relock them in
3184 order. Ensure that we don't try to re-lock the same resource.
3187 Frequency : only very large systems with liblustre clients
3189 Description: slow eviction of liblustre clients with the "evict_by_nid" RPC
3190 Details : Use asynchronous set_info RPCs to send the "evict_by_nid" to
3191 all OSTs in parallel. This allows the eviction of stale liblustre
3192 clients to proceed much faster than if they were done in series,
3193 and also offers similar improvements for other set_info RPCs.
3198 Description: excessive CPU usage during initial read phase on client
3199 Details : During the initial read phase on a client, it would agressively
3200 retry readahead on the file, consuming too much CPU and impacting
3201 performance (since 1.4.5.8). Improve the readahead algorithm
3202 to avoid this, and also improve some other common cases (read
3203 of small files in particular, where "small" is files smaller than
3204 /proc/fs/lustre/llite/*/max_read_ahead_whole_mb, 2MB by default).
3209 Description: MDS crash when receiving packet with unknown intent.
3210 Details : Do not LBUG in unknown intent case, just return -EFAULT
3212 Severity : enhancement
3213 Bugzilla : 9293, 9385
3214 Description: MDS RPCs are serialised on client. This is unnecessary for some.
3215 Details : Do not serialize getattr (non-intent version) and statfs.
3218 Frequency : occasional, when OST network is overloaded/intermittent
3220 Description: client evicted by OST after bulk IO timeout
3221 Details : If a client sends a bulk IO request (read or write) the OST
3222 may evict the client if it is unresposive to its data GET/PUT
3223 request. This is incorrect if the network is overloaded (takes
3224 too long to transfer the RPC data) or dropped the OST GET/PUT
3225 request. There is no need to evict the client at all, since
3226 the pinger and/or lock callbacks will handle this, and the
3227 client can restart the bulk request.
3230 Frequency : Always when mmapping file with no objects
3232 Description: client crashes when mmapping file with no objects
3233 Details : Check that we actually have objects in a file before doing any
3234 operations on objects in ll_vm_open, ll_vm_close and
3240 Description: Request leak when working with deleted CWD
3241 Details : Introduce advanced request refcount tracking for requests
3242 referenced from lustre intent.
3244 Severity : Enhancement
3246 Description: Cache open file handles on client.
3247 Details : MDS now will return special lock along with openhandle, if
3248 requested and client is allowed to hold openhandle, even if unused,
3249 until such a lock is revoked. Helps NFS a lot, since NFS is opening
3250 closing files for every read/write openration.
3252 Severity : Enhancement
3254 Description: Cache open negative dentries on client when possible.
3255 Details : Guard negative dentries with UPDATE lock on parent dir, drop
3256 negative dentries on lock revocation.
3261 Description: Remounting a client read-only wasn't possible with a zconf mount
3262 Details : It wasn't possible to remount a client read-only with llmount.
3264 Severity : enhancement
3265 Description: Include MPICH 1.2.6 Lustre ADIO interface patch
3266 Details : In lustre/contrib/ or /usr/share/lustre in RPM a patch for
3267 MPICH is included to add Lustre-specific ADIO interfaces.
3268 This is based closely on the UFS ADIO layer and only differs
3269 in file creation, in order to allow the OST striping to be set.
3270 This is user-contributed code and not supported by CFS.
3275 Description: extended inode attributes (immutable, append-only) work improperly
3276 when 2.4 and 2.6 kernels are used on client/server or vice versa
3277 Details : Introduce kernel-independent values for these flags.
3279 Severity : enhancement
3282 Description: Allow fractional MB tunings for lustre in /proc/ filesystem.
3283 Details : Many of the /proc/ tunables can only be tuned at a megabyte
3284 granularity. Now, Fractional MB granularity is be supported,
3285 this is very useful for low memory system.
3287 Severity : enhancement
3289 Description: Getattr by fid
3290 Details : Getting a file attributes by its fid, obtaining UPDATE|LOOKUP
3291 locks, avoids extra getattr rpc requests to MDS, allows '/' to
3292 have locks and avoids getattr rpc requests for it on every stat.
3295 Frequency : Always, for filesystems larger than 2TB
3297 Description: ldiskfs crash at mount for filesystem larger than 2TB with mballoc
3298 Details : Kenrel kmalloc limits allocations to 128kB and this prevents
3299 filesystems larger than 2TB to be mounted with mballoc enabled.
3302 Frequency : Always, for 32-bit kernel without CONFIG_LBD and filesystem > 2TB
3304 Description: ldiskfs crash at mount for filesystem larger than 2TB with mballoc
3305 Details : If a 32-bit kernel is compiled without CONFIG_LBD enabled and a
3306 filesystems larger than 2TB is mounted then the kernel will
3307 silently corrupt the start of the filesystem. CONFIG_LBD is
3308 enabled for all CFS-supported kernels, but the possibility of
3309 this happening with a modified kernel config exists.
3311 Severity : enhancement
3313 Description: add client O_DIRECT support for 2.6 kernels
3314 Details : It is now possible to do O_DIRECT reads and writes to files
3315 in the Lustre client mountpoint on 2.6 kernel clients.
3317 Severity : enhancement
3319 Description: parallel glimpse, setattr, statfs, punch, destroy requests
3320 Details : Sends glimpse, setattr, statfs, punch, destroy requests to OSTs in
3321 parallel, not waiting for response from every OST before sending
3322 a rpc to the next OST.
3327 Description: setattr vs write race when updating file timestamps
3328 Details : Client processes that update a file timestamp into the past
3329 right after writing to the file (e.g. tar) it is possible that
3330 the updated file modification time can be reset to the current
3331 time due to a race between processing the setattr and write RPC.
3333 Severity : enhancement
3335 Description: Bring 'lfs find' closer in line with regular Linux find.
3336 Details : lfs find util supports -atime, -mtime, -ctime, -maxdepth, -print,
3337 -print0 options and obtains all the needed info through the lustre
3340 Severity : enhancement
3342 Description: support up to 1024 configured devices on one node
3343 Details : change obd_dev array from statically allocated to dynamically
3344 allocated structs as they are first used to reduce memory usage
3349 Description: Flush dirty partially truncated pages during truncate
3350 Details : Immediatelly flush partially truncated pages in filter_setattr,
3351 this way we completely avoid having any pages in page cache on OST
3352 and can retire ugly workarounds during writes to flush such pages.
3357 Description: i_sem vs transaction deadlock in mds_obd_destroy during unlink.
3358 Details : protect inode from truncation within vfs_unlink() context
3359 just take a reference before calling vfs_unlink() and release it
3360 when parent's i_sem is free.
3365 Description: last_id value checked outside lock on OST caused LASSERT failure
3366 Details : If there were multiple MDS->OST object precreate requests in
3367 flight, it was possible that the OST's last object id was checked
3368 outside a lock and incorrectly tripped an assertion. Move checks
3369 inside locks, and discard old precreate requests.
3372 Frequency : always, if extents are used on OSTs
3374 Description: index ei_leaf_hi (48-bit extension) is not zeroed in extent index
3375 Details : OSTs using the extents format would not zero the high 16 bits of
3376 the index physical block number. This is not a problem for any
3377 OST filesystems smaller than 16TB, and no kernels support ext3
3378 filesystems larger than 16TB yet. This is fixed in 1.4.7 (all
3379 new/modified files) and can be fixed for existing filesystems
3380 with e2fsprogs-1.39-cfs1.
3385 Description: import connection selection may be incorrect if timer wraps
3386 Details : Using a 32-bit jiffies timer with HZ=1000 may cause backup
3387 import connections to be ignored if the 32-bit jiffies counter
3388 wraps. Use a 64-bit jiffies counter.
3391 Frequency : very large clusters immediately after boot
3393 Description: LNET request buffers exhausted under heavy short-term load
3394 Details : If a large number of client requests are generated on a service
3395 that has previously never seen so many requests it is possible
3396 that the request buffer growth cannot keep up with the spike in
3397 demand. Instead of dropping incoming requests, they are held in
3398 the LND until the RPC service can accept more requests.
3401 Frequency : Sometimes during replay
3403 Description: Assertion failure in ll_local_open after replay.
3404 Details : If replay happened on an open request reply before we were able
3405 to set replay handler, reply will become not swabbed tripping the
3406 assertion in ll_local_open. Now we set the handler right after
3407 recognising of open request
3410 Frequency : very rare
3412 Description: kernel reports "badness in vsnprintf"
3413 Details : Reading from the "recovery_status" /proc file in small chunks
3414 may cause a negative length in lprocfs_obd_rd_recovery_status()
3415 call to vsnprintf() (which is otherwise harmless). Exit early
3416 if there is no more space in the output buffer.
3418 Severity : enhancement
3420 Description: clear OBD RPC statistics by writing to them
3421 Details : It is now possible to clear the OBD RPC statistics by writing
3422 to the "stats" file.
3427 Description: Client mtime is not the same on different clients after utimes
3428 Details : In some cases, the client was using the utimes() syscall on
3429 a file cached on another node. The clients now validate the
3430 ctime from the MDS + OSTs to determine which one is right.
3435 Description: Inability to activate failout mode
3436 Details : lconf script incorrectly assumed that in pythong string's numeric
3437 value is used in comparisons.
3440 Frequency : always with multiple stripes per file
3442 Description: Inefficient object allocation for mutli-stripe files
3443 Details : When selecting which OSTs to stripe files over, for files with
3444 a stripe count that divides evenly into the number of OSTs,
3445 the MDS is always picking the same starting OST for each file.
3446 Return the OST selection heuristic to the original design.
3451 Description: mount failures may take full timeout to return an error
3452 Details : Under some heavy load conditions it is possible that a
3453 failed mount can wait for the full obd_timeout interval,
3454 possibly several minutes, before reporting an error.
3455 Instead return an error as soon as the status is known.
3457 Frequency : quota enabled and large files being deleted
3459 Description: releasing more than 4GB of quota at once hangs OST
3460 Details : If a user deletes more than 4GB of files on a single OST it
3461 will cause the OST to spin in an infinite loop. Release
3462 quota in < 4GB chunks, or use a 64-bit value for 1.4.7.1+.
3467 Description: statfs data retrieved from /proc may be stale or zero
3468 Details : When reading per-device statfs data from /proc, in the
3469 {kbytes,files}_{total,free,avail} files, it may appear
3470 as zero or be out of date.
3473 Frequency : systems with MD RAID1 external journal devices
3475 Description: lconf's call to blkid is confused by RAID1 journal devices
3476 Details : Use the "blkid -l" flag to locate the MD RAID device instead
3477 of returning all block devices that match the journal UUID.
3480 Frequency : always, for aggregate stripe size over 4GB
3482 Description: assertion fails when trying to use 4GB stripe size
3483 Details : Use "setstripe" to set stripe size over 4GB will fail the kernel,
3484 complaining "ASSERTION(lsm->lsm_xfersize != 0)"
3487 Frequency : always on ppc64
3489 Description: the first write on an ext3 filesystem with mballoc got stuck
3490 Details : ext3_mb_generate_buddy() uses find_next_bit() which does not
3491 perform endianness conversion.
3493 ------------------------------------------------------------------------------
3495 02-14-2006 Cluster File Systems, Inc. <info@clusterfs.com>
3497 * WIRE PROTOCOL CHANGE. This version of Lustre networking WILL NOT
3498 INTEROPERATE with older versions automatically. Please read the
3499 user documentation before upgrading any part of a live system.
3500 * WARNING: Lustre networking configuration changes are required with
3501 this release. See https://bugzilla.clusterfs.com/show_bug.cgi?id=10052
3504 * Support for kernels:
3505 2.6.9-22.0.2.EL (RHEL 4)
3506 2.6.5-7.244 (SLES 9)
3507 2.6.12.6 vanilla (kernel.org)
3510 Severity : enhancement
3511 Bugzilla : 7981/8208
3512 Description: Introduced Lustre Networking (LNET)
3513 Details : LNET is new networking infrastructure for Lustre, it includes
3514 a reorganized network configuration mode (see the user
3515 documentation for full details) as well as support for routing
3516 between different network fabrics. Lustre Networking Devices
3517 (LNDs) for the supported network fabrics have also been
3518 created for this new infrastructure.
3520 Severity : enhancement
3521 Description: Introduced Access control lists
3522 Details : clients can set ACLs on files and directories in order to have
3523 more fine-grained permissions than the standard Unix UGO+RWX.
3524 The MDS must be started with the "-o acl" mount option.
3526 Severity : enhancement
3527 Description: Introduced filesystem quotas
3528 Details : Administrators may now establish per-user quotas on the
3531 Severity : enhancement
3533 Description: Configuration change for the XT3
3534 The PTLLND is now used to run Lustre over Portals on the XT3
3535 The configure option(s) --with-cray-portals are no longer used.
3536 Rather --with-portals=<path-to-portals-includes> is used to
3537 enable building on the XT3. In addition to enable XT3 specific
3538 features the option --enable-cray-xt3 must be used.
3543 Description: Running on many-way SMP OSTs can trigger oops in llcd_send()
3544 Details : A race between allocating a new llcd and re-getting the llcd_lock
3545 allowed another thread to grab newly-allocated llcd.
3547 Severity : enhancement
3549 Description: 2.6 OST async journal commit and locking fix to improve performance
3550 Details : The filter_direct_io()+filter_commitrw_write() journal commits for
3551 2.6 kernels are now async as they already were in 2.4 kernels so
3552 that they can commit concurrently with the network bulk transfer.
3553 For block-allocated files the filter allocation semaphore is held
3554 to avoid filesystem fragmentation during allocation. BKL lock
3555 removed for 2.6 xattr operations where it is no longer needed.
3560 Description: lconf incorrectly determined whether two IP networks could talk
3561 Details : In some more complicated routing and multiple-network
3562 configurations, lconf will avoid trying to make a network
3563 connection to a disjoint part of the IP space. It was doing the
3564 math incorrectly for one set of cases.
3569 Description: Fix for potential infinite loop processing records in an llog.
3570 Details : If an llog record is corrupted/zeroed, it is possible to loop
3571 forever in llog_process(). Validate the llog record length
3572 and skip the remainder of the block on error.
3575 Frequency : occasional (liblustre only)
3577 Description: liblustre could not open files whose last component is a symlink
3578 Details : sysio_path_walk() would incorrectly pass the open intent to
3579 intermediate path components.
3582 Frequency : rare (liblustre only with non-standard tuning)
3583 Bugzilla : 7201 (7350)
3584 Description: Tuning the MDC DLM LRU size to zero triggers client LASSERT
3585 Details : llu_lookup_finish_locks() tries to set lock data on a lock
3586 after it has been released, only do this for referenced locks
3588 Severity : enhancement
3590 Description: specifying an (invalid) directory default stripe_size of -1
3591 would reset the directory default striping
3592 Details : stripe_size -1 was used internally to signal directory stripe
3593 removal, now use "all default" to signal dir stripe removal
3594 as a directory striping of "all default" is not useful
3597 Frequency : common for large clusters running liblustre clients
3599 Description: doing an ls when liblustre clients are running is slow
3600 Details : sending a glimpse AST to a liblustre client waits for every AST
3601 to time out, as liblustre clients will not respond. Since they
3602 cannot cache data we refresh the OST lock LVB from disk instead.
3604 Severity : enhancement
3606 Description: doing an ls at the same time as file IO can be slow
3607 Details : enqueue and other "small" requests can be blocked behind many
3608 large IO requests. Create a new OST IO portal for non-IO
3609 requests so they can be processed faster.
3612 Frequency : rare (only HPUX clients mounting unsupported re-exported NFS vol)
3614 Description: an HPUX NFS client would get -EACCESS when ftruncate()ing a newly
3615 created file with mode 000
3616 Details : the Linux NFS server relies on an MDS_OPEN_OWNEROVERRIDE hack to
3617 allow an ftruncate() as a non-root user to a file with mode 000.
3618 Lustre now respects this flag to disable mode checks when
3619 truncating a file owned by the user
3622 Frequency : liblustre-only, when liblustre client dies unexpectedly or becomes
3625 Description: Revoking locks from clients that went dead or catatonic might take
3627 Details : New lock flags FL_CANCEL_ON_BLOCK used by liblustre makes
3628 cancellation of such locks instant on servers without waiting for
3629 any reply from clients. Clients drops these locks when cancel
3630 notification from server is received without replying.
3633 Frequency : liblustre-only, when liblustre client dies or becomes busy
3635 Description: Doing ls on Linux clients can take a long time with active
3637 Details : Liblustre client cannot handle ASTs in timely manner, so avoid
3638 granting such locks to it in the first place if possible. Locks
3639 are taken by proxy on the OST during the read or write and
3640 dropped immediately afterward. Add connect flags handling, do
3641 not grant locks to liblustre clients for glimpse ASTs.
3643 Severity : enhancement
3645 Description: Improve read-ahead algorithm to avoid excessive IO for random reads
3646 Details : Existing read-ahead algorithm is tuned for the case of streamlined
3647 sequential reads and behaves badly with applications doing random
3648 reads. Improve it by reading ahead at least read region, and
3649 avoiding excessive large RPC for small reads.
3651 Severity : enhancement
3653 Description: Creating more than 1000 files for a single job may cause a load
3654 imbalance on the OSTs if there are also a large number of OSTs.
3655 Details : qos_prep_create() uses an OST index reseed value that is an
3656 even multiple of the number of available OSTs so that if the
3657 reseed happens in the middle of the object allocation it will
3658 still utilize the OSTs as uniformly as possible.
3663 Description: OST or MDS may oops in ping_evictor_main()
3664 Details : ping_evictor_main() drops obd_dev_lock if deleting a stale export
3665 but doesn't restart at beginning of obd_exports_timed list
3668 Severity : enhancement
3670 Description: improve by-nid export eviction on the MDS and OST
3671 Details : allow multiple exports with the same NID to be evicted at one
3672 time without re-searching the exports list.
3675 Frequency : rare, only with supplementary groups enabled on SMP 2.6 kernels
3677 Description: MDS may oops in groups_free()
3678 Details : in rare race conditions a newly allocated group_info struct is
3679 freed again, and this can be NULL. The 2.4 compatibility code
3680 for groups_free() checked for a NULL pointer, but 2.6 did not.
3683 Frequency : common for liblustre clients doing little filesystem IO
3684 Bugzilla : 9352, 7313
3685 Description: server may evict liblustre clients accessing contended locks
3686 Details : if a client is granted a lock or receives a completion AST
3687 with a blocking AST already set it would not reply to the AST
3688 for LDLM_FL_CANCEL_ON_BLOCK locks. It now replies to such ASTs.
3691 Frequency : lfs setstripe, only systems with more than 160 OSTs
3693 Description: unable to set striping with a starting offset beyond OST 160
3694 Details : llapi_create_file() incorrectly limited the starting stripe
3695 index to the maximum single-file stripe count.
3698 Frequency : LDAP users only
3700 Description: lconf did not handle in-kernel recovery with LDAP properly
3701 Details : lconf/LustreDB get_refs() is searching the wrong namespace
3703 Severity : enhancement
3705 Description: bind OST threads to NUMA nodes to improve performance
3706 Details : all OST threads are uniformly bound to CPUs on a single NUMA
3707 node and do their allocations there to localize memory access
3709 Severity : enhancement
3711 Description: llmount can determine client NID directly from Myrinet (GM)
3712 Details : the client NID code from gmnalnid was moved directly into
3713 llmount, removing the need to use this or specifying the
3714 client NID explicitly when mounting GM clients with zeroconf
3717 Frequency : if client is started with down MDS
3719 Description: if client is started with down MDS mount hangs in ptlrpc_queue_wait
3720 Details : Having an LWI_INTR() wait event (interruptible, but no timeout)
3721 will wait indefinitely in ptlrpc_queue_wait->l_wait_event() after
3722 ptlrpc_import_delayed_req() because we didn't check if the
3723 request was interrupted, and we also didn't break out of the
3724 event loop if there was no timeout
3729 Description: data loss during non-page-aligned writes to a single file from
3730 both multiple nodes and multiple threads on one node at same time
3731 Details : updates to KMS and lsm weren't protected by common lock. Resulting
3732 inconsistency led to false short-reads, that were cached and later
3733 used by ->prepare_write() to fill in partially written page,
3734 leading to data loss.
3737 Frequency : always, if lconf --abort_recovery used
3739 Description: lconf --abort_recovery fails with 'Operation not supported'
3740 Details : lconf was attempting to abort recovery on the MDT device and not
3743 Severity : enhancement
3745 Description: remove cleanup logs
3746 Details : replace lconf-generated cleanup logs with lustre internal
3747 cleanup routines. Eliminates the need for client-cleanup and
3750 Severity : enhancement
3752 Description: add support for EAs (user and system) on lustre filesystems
3753 Details : it is now possible to store extended attributes in the Lustre
3754 client filesystem, and with the user_xattr mount option it
3755 is possible to allow users to store EAs on their files also
3757 Severity : enhancement
3759 Description: Add possibility (config option) to show minimal available OST free
3761 Details : When compiled with --enable-mindf configure option, statfs(2)
3762 (and so, df) will return least minimal free space available from
3763 all OSTs as amount of free space on FS, instead of summary of
3764 free spaces of all OSTs.
3766 Severity : enhancement
3768 Description: do not expand extent locks acquired on OST-side
3769 Details : Modify ldlm_extent_policy() to not expand local locks, acquired
3770 by server: they are not cached anyway.
3773 Frequency : when mmap is used/binaries executed from Lustre
3775 Description: Unmmap pages before throwing them away from read cache.
3776 Details : llap_shrink cache now attempts to unmap pages before discarding
3777 them (if unmapping failed - do not discard). SLES9 kernel has
3778 extra checks that trigger if this unmapping is not done first.
3783 Description: lconf didn't resolve symlinks before checking to see whether a
3784 given mountpoint was already in use
3787 Frequency : when migrating failover services
3788 Bugzilla : 6395, 9514
3789 Description: When migrating a subset of services from a node (e.g. failback
3790 from a failover service node) the remaining services would
3791 time out and evict clients.
3792 Details : lconf --force (implied by --failover) sets the global obd_timeout
3793 to 5 seconds in order to quickly disconnect, but this caused
3794 other RPCs to time out too quickly. Do not change the global
3795 obd_timeout for force cleanup, only set it for DISCONNECT RPCs.
3797 Severity : enhancement
3798 Frequency : if MDS is started with down OST
3799 Bugzilla : 9439,5706
3800 Description: Allow startup/shutdown of an MDS without depending on the
3801 availability of the OSTs.
3802 Details : Asynchronously call mds_lov_synchronize during MDS startup.
3803 Add appropriate locking and lov-osc refcounts for safe
3804 cleaning. Add osc abort_inflight calls in case the
3805 synchronize never started.
3808 Frequency : occasional (Cray XT3 only)
3810 Description: root not authorized to access files in CRAY_PORTALS environment
3811 Details : The client process capabilities were not honoured on the MDS in
3812 a CRAY_PORTALS/CRAY_XT3 environment. If the file had previously
3813 been accessed by an authorized user then root was able to access
3814 the file on the local client also. The root user capabilities
3815 are now allowed on the MDS, as this environment has secure UID.
3818 Frequency : occasional
3820 Description: ldiskfs "too long searching" message happens too often
3821 Details : A debugging message (otherwise harmless) prints too often on
3822 the OST console. This has been reduced to only happen when
3823 there are fragmentation problems on the filesystem.
3828 Description: Division by zero in statfs when all OSCs are inactive
3829 Details : lov_get_stripecnt() returns zero due to incorrect order of checks,
3830 lov_statfs divides by value returned by lov_get_stripecnt().
3834 Bugzilla : 9489, 3273
3835 Description: First write from each client to each OST was only 4kB in size,
3836 to initialize client writeback cache, which caused sub-optimal
3837 RPCs and poor layout on disk for the first writen file.
3838 Details : Clients now request an initial cache grant at (re)connect time
3839 and so that they can start streaming writes to the cache right
3840 away and always do full-sized RPCs if there is enough data.
3841 If the OST is rebooted the client also re-establishes its grant
3842 so that client cached writes will be honoured under the grant.
3847 Description: Slow ls (and stat(2) syscall) on files residing on IO-loaded OSTs
3848 Details : Now I/O RPCs go to different portal number and (presumably) fast
3849 lock requests (and glimses) and other RPCs get their own service
3850 threads pool that should be able to service those RPCs
3853 Severity : enhancement
3855 Description: Ability to exchange lustre version between client and servers and
3856 issue warnings at client side if client is too old. Also for
3857 liblustre clients there is ability to refuse connection of too old
3859 Details : New 'version' field is added to connect data structure that is
3860 filled with version info. That info is later checked by server and
3864 Frequency : rare, liblustre only.
3865 Bugzilla : 9296, 9581
3866 Description: Two simultaneous writes from liblustre at offset within same page
3867 might proceed at the same time overwriting eachother with stale
3869 Details : I/O lock withing llu_file_prwv was released too early, before data
3870 actually was hitting the wire. Extended lock-holding time until
3871 server acknowledges receiving data.
3874 Frequency : extremely rare. Never observed in practice.
3876 Description: avoid generating lustre_handle cookie of 0.
3877 Details : class_handle_hash() generates handle cookies by incrementing
3878 global counter, and can hit 0 occasionaly (this is unlikely, but
3879 not impossible, because initial value of cookie counter is
3880 selected randonly). Value of 0 is used as a sentinel meaning
3881 "unassigned handle" --- avoid it. Also coalesce two critical
3882 sections in this function into one.
3884 Severity : enhancement
3886 Description: allow liblustre clients to delegate truncate locking to OST
3887 Details : To avoid overhead of locking, liblustre client instructs OST to
3888 take extent lock in ost_punch() on client's behalf. New connection
3889 flag is added to handle backward compatibility.
3891 Severity : enhancement
3892 Bugzilla : 4928, 7341, 9758
3893 Description: allow number of OST service threads to be specified
3894 Details : a module parameter allows the number of OST service threads
3895 to be specified via "options ost ost_num_threads={N}" in the
3896 OSS's /etc/modules.conf or /etc/modprobe.conf.
3900 Bugzilla : 6146, 9635, 9895
3901 Description: servers crash with bad pointer in target_handle_connect()
3902 Details : In rare cases when a client is reconnecting it was possible that
3903 the connection request was the last reference for that export.
3904 We would temporarily drop the export reference and get a new
3905 one, but this may have been the last reference and the export
3906 was just destroyed. Get new reference before dropping old one.
3908 Severity : enhancement
3909 Frequency : if client is started with failover MDS
3911 Description: Allow multiple MDS hostnames in the mount command
3912 Details : Try to read the configuration from all specified MDS
3913 hostnames during a client mount in case the "primary"
3916 Severity : enhancement
3918 Description: Stop sending data to evicted clients as soon as possible.
3919 Details : Check if the client we are about to send or are sending data to
3920 was evicted already. (Check is done every second of waiting,
3921 for which l_wait_event interface was extended to allow checking
3922 of exit condition at specified intervals).
3925 Frequency : rare, normally only when NFS exporting is done from client
3927 Description: 'bad disk LOV MAGIC: 0x00000000' error when chown'ing files
3929 Details : Make mds_get_md() recognise empty md case and set lmm size to 0.
3932 Frequency : always, if srand() is called before liblustre initialization
3934 Description: Liblustre uses system PRNG disturbing its usage by user application
3935 Details : Introduce internal to lustre fast and high-quality PRNG for
3936 lustre usage and make liblustre and some other places in generic
3937 lustre code to use it.
3939 Severity : enhancement
3940 Bugzilla : 9477, 9557, 9870
3941 Description: Verify that the MDS configuration logs are updated when xml is
3942 Details : Check if the .xml configuration logs are newer than the config
3943 logs stored on the MDS and report an error if this is the case.
3944 Request --write-conf, or allow starting with --old_conf.
3946 Severity : enhancement
3948 Description: Handle symlinks in the path when checking if Lustre is mounted.
3949 Details : Resolve intermediate symlinks when checking if a client has
3950 mounted a filesystem to avoid duplicate client mounts.
3955 Description: lconf can hit an error exception but still return success.
3956 Details : The lconf command catches the Command error exception at the top
3957 level script context and will exit with the associated exit
3958 status, but doesn't ensure that this exit status is non-zero.
3963 Description: failure of ptlrpc thread startup can cause oops
3964 Details : Starting a ptlrpc service thread can fail if there are a large
3965 number of threads or the server memory is very fragmented.
3966 Handle this without oopsing.
3969 Frequency : always, only if liblustre and non-default acceptor port was used
3971 Description: liblustre cannot connect to servers with non-default acceptor port
3972 Details : tcpnal_set_default_params() was not called and was therefore
3973 ignoring the environment varaible TCPNAL_PORT, as well as other
3974 TCPNAL_ environment variables
3979 Description: two objects could be created on the same OST for a single file
3980 Details : If an OST is down, in some cases it was possible to create two
3981 objects on a single OST for a single file. No problems other
3982 than potential performance impact and spurious error messages.
3986 Bugzilla : 5681, 9562
3987 Description: Client may oops in ll_unhash_aliases
3988 Details : Client dcache may become inconsistent in race condition.
3989 In some cases "getcwd" can fail if the current directory is
3995 Description: Inode refcounting problems in NFS export code
3996 Details : link_raw functions used to call d_instantiate without obtaining
3997 extra inode reference first.
4001 Bugzilla : 9942, 9903
4002 Description: Referencing freed requests leading to crash, memleaks with NFS.
4003 Details : We used to require that call to ll_revalidate_it was always
4004 followed by ll_lookup_it. Also with revalidate_special() it is
4005 possible to call ll_revalidate_it() twice for the same dentry
4006 even if first occurence returned success. This fix changes semantic
4007 between DISP_ENQ_COMPLETE disposition flag to mean there is extra
4008 reference on a request referred from the intent.
4009 ll_intent_release() then releases such a request.
4012 Frequency : rare, normally benchmark loads only
4014 Description: unlinked inodes were kept in memory on the client
4015 Details : If a client is repeatedly creating and unlinking files it
4016 can accumulate a lot of stale inodes in the inode slab cache.
4017 If there is no other client load running this can cause the
4018 client node to run out of memory. Instead flush old inodes
4019 from client cache that have the same inode number as a new inode.
4022 Frequency : SLES9 2.6.5 kernel and long filenames only
4023 Bugzilla : 9969, 10379
4024 Description: utime reports stale NFS file handle
4025 Details : SLES9 uses out-of-dentry names in some cases, which confused
4026 the lustre dentry revalidation. Change it to always use the
4030 Frequency : rare, unless heavy write-truncate concurrency is continuous
4031 Bugzilla : 4180, 6984, 7171, 9963, 9331
4032 Description: OST becomes very slow and/or deadlocked during object unlink
4033 Details : filter_destroy() was holding onto the parent directory lock
4034 while truncating+unlinking objects. For very large objects this
4035 may block other threads for a long time and slow overall OST
4036 responsiveness. It may also be possible to get a lock ordering
4037 deadlock in this case, or run out of journal credits because of
4038 the combined truncate+unlink. Solution is to do object truncate
4039 first in one transaction without parent lock, and then do the
4040 final unlink in a new transaction with the parent lock. This
4041 reduces the lock hold time dramatically.
4044 Frequency : rare, 2.4 kernels only
4046 Description: MDS or OST cleanup may trip kernel BUG when dropping kernel lock
4047 Details : mds_cleanup() and filter_cleanup() need to drop the kernel lock
4048 before unmounting their filesystem in order to avoid deadlock.
4049 The kernel_locked() function in 2.4 kernels only checks whether
4050 the kernel lock is held, not whether it is this process that is
4051 holding it as 2.6 kernels do.
4056 Description: MDS or OST may oops/LBUG if a client is connecting multiple times
4057 Details : The client ptlrpc code may be trying to reconnect to a down
4058 server before a previous connection attempt has timed out.
4059 Increase the reconnect interval to be longer than the connection
4060 timeout interval to avoid sending duplicate connections to
4064 Frequency : echo_client brw_test command
4066 Description: fix echo_client to work with OST preallocated code
4067 Details : OST preallocation code (5137) didn't take echo_client IO path
4068 into account: echo_client calls filter methods outside of any
4069 OST thread and, hence, there is no per-thread preallocated
4070 pages and buffers to use. Solution: hijack pga pages for IO. As
4071 a byproduct, this avoids unnecessary data copying.
4075 Bugzilla : 3555, 5962, 6025, 6155, 6296, 9574
4076 Description: Client can oops in mdc_commit_close() after open replay
4077 Details : It was possible for the MDS to return an open request with no
4078 transaction number in mds_finish_transno() if the client was
4079 evicted, but without actually returning an error. Clients
4080 would later try to replay that open and may trip an assertion
4081 Simplify the client close codepath, and always return an error
4082 from the MDS in case the open is not successful.
4085 Frequency : rare, 2.6 OSTs only
4087 Description: OST may deadlock under high load on fragmented files
4088 Details : If there was a heavy load and highly-fragmented OST filesystems
4089 it was possible to have all the OST threads deadlock waiting on
4090 allocation of biovecs, because the biovecs were not released
4091 until the entire RPC IO was completed. Instead, release biovecs
4092 as soon as they are complete to ensure forward IO progress.
4094 Severity : enhancement
4096 Description: Support for specifying external journal device at mount
4097 Details : If an OST or MDS device is formatted with an external journal
4098 device, this device major/minor is stored in the ext3 superblock
4099 and may not be valid for failover. Allow detecting and
4100 specifying the external journal at mount time.
4105 Description: Mounting an MDS with pending unlinked files may cause oops
4106 Details : target_finish_recovery() calls mds_postrecov() which returned
4107 the number of orphans unlinked. mds_lov_connect->mds_postsetup()
4108 considers this an error and immediately begins cleaning up the
4109 lov, just after starting the mds_lov process
4111 Severity : enhancement
4113 Description: Implement 'lfs df' to report actual free space on per-OST basis
4114 Details : Add sub-command 'df' on 'lfs' to report the disk space usage of
4115 MDS/OSDs. Usage: lfs df [-i][-h]. Command Options: '-i' to report
4116 usage of objects; '-h' to report in human readable format.
4118 ------------------------------------------------------------------------------
4120 08-26-2005 Cluster File Systems, Inc. <info@clusterfs.com>
4127 Description: Mounting an ldiskfs file system with mballoc may crash OST node.
4128 Details : ldiskfs mballoc code may reference an uninitialized buddy struct
4129 at startup during orphan unlinking. Instead, skip buddy update
4130 before setup, as it will be regenerated after recovery is complete.
4135 Description: If an OST is inactive, its locks might reference stale inodes.
4136 Details : lov_change_cbdata() must iterate over all namespaces, even if
4137 they are inactive to clear inode references from the lock.
4139 Severity : enhancement
4140 Frequency : occasional, if non-standard max_dirty_mb used
4142 Description: Client will block write RPCs if not enough grant
4143 Details : If a client has max_dirty_mb smaller than max_rpcs_in_flight,
4144 then the client will block writes while waiting for another RPC
4145 to complete instead of consuming its dirty limit. With change
4146 we get improved performance when max_dirty_mb is small.
4148 Severity : enhancement
4149 Bugzilla : 3389, 6253
4150 Description: Add support for supplementary groups on the MDS.
4151 Details : The MDS has an upcall /proc/fs/lustre/mds/{mds}/group_upcall
4152 (set to /usr/sbin/l_getgroups if enabled) which will do MDS-side
4153 lookups for user supplementary groups into a cache.
4157 Description: O_CREAT|O_EXCL open flags in liblustre always return -EEXIST
4158 Details : Make libsysio to not enforce O_EXCL by clearing the flag,
4159 for liblustre O_EXCL is enforced by MDS.
4163 Description: readdir never returns NULL in liblustre.
4164 Details : Corrected llu_iop_getdirentries logic, to return offset of next
4165 dentry in struct dirent.
4169 Frequency : liblustre only, depends on application IO pattern
4170 Description: liblustre clients evicted if not contacting servers
4171 Details : Don't put liblustre clients into the ping_evictor list, so
4172 they will not be evicted by the pinger ever.
4174 Severity : enhancement
4176 Description: Add ability to evict clients by NID from MDS.
4177 Details : By echoing "nid:$NID" string into
4178 /proc/fs/lustre/mds/.../evict_client client with nid that equals to
4179 $NID would be instantly evicted from this MDS and from all active
4180 OSTs connected to it.
4184 Description: Do not query file size twice, somewhat slowing stat(2) calls.
4185 Details : lookup_it_finish() used to query file size from OSTs that was not
4190 Description: service threads change working directory to that of init
4191 Details : Starting lustre service threads may pin the working directory
4192 of the parent thread, making that filesystem busy. Threads
4193 now change to the working directory of init to avoid this.
4197 Frequency : during shutdown only
4198 Description: shutdown with a failed MDS or OST can cause unmount to hang
4199 Details : Don't resend DISCONNECT messages in ptlrpc_disconnect_import()
4204 Frequency : 2.6 only
4205 Description: chmod/chown may include an extra supplementary group
4206 Details : ll{,u}_mdc_pack_op_data() does not properly initialize the
4207 supplementary group and if none is specified this is used.
4210 Bugzilla : 5479 (6816)
4212 Description: Racing open + rm can assert client in mdc_set_open_replay_data()
4213 Details : If lookup is in progress on a file that is unlinked we might try
4214 to revalidate the inode and fail in revalidate after lookup is
4215 complete and ll_file_open() enqueues the open again but
4216 it_open_error() was not checking DISP_OPEN_OPEN errors correctly.
4219 Frequency : always, if lconf --abort_recovery used
4221 Description: lconf --abort_recovery fails with 'Operation not supported'
4222 Details : lconf was attempting to abort recovery on the MDT device and not
4225 ------------------------------------------------------------------------------
4227 2005-08-08 Cluster File Systems, Inc. <info@clusterfs.com>
4232 Frequency : rare (only unsupported configurations with a node running as an
4234 Bugzilla : 6514, 5137
4235 Description: Mounting a Lustre file system on a node running as an OST could
4237 Details : OSTs now preallocates memory needed to write out data at
4238 startup, instead of when needed, to avoid having to
4239 allocate memory in possibly low memory situations.
4240 Specifically, if the file system is mounted on on OST,
4241 memory pressure could force it to try to write out data,
4242 which it needed to allocate memory to do. Due to the low
4243 memory, it would be unable to do so and the node would
4244 become unresponsive.
4246 Severity : enhancement
4248 Description: Addition of lconf --service command line option
4249 Details : lconf now accepts a '--service <arg>' option, which is
4250 shorthand for 'lconf --group <arg> --select <arg>=<hostname>'
4252 Severity : enhancement
4254 Description: Failover mode is now the default for OSTs.
4255 Details : By default, OSTs will now run in failover mode. To return to
4256 the old behaviour, add '--failout' to the lmc line for OSTs.
4258 Severity : enhancement
4260 Description: Health checks are now provided for MDS and OSTs
4261 Details : Additional detailed health check information on MSD and OSTs
4262 is now provided through the procfs health_check value.
4265 Frequency : occasional, depends on IO load
4267 Description: Disk fragmentation on the OSTs could eventually cause slowdowns
4268 after numerous create/delete cycles
4269 Details : The ext3 inode allocation policy would not allocate new inodes
4270 very well on the OSTs because there are no new directories
4271 being created. Instead we look for groups with free space if
4272 the parent directories are nearly full.
4277 Description: Network or server problems during mount may cause partially
4278 mounted clients instead of returning an error.
4279 Details : The config llog parsing code may overwrite the error return
4280 code during mount error handling, returning success instead
4286 Description: MDS can fail to allocate large reply buffers
4287 Details : After long uptimes the MDS can fail to allocate large reply
4288 buffers (e.g. zconf client mount config records) due to memory
4289 fragmentation or consumption by the buffer cache. Preallocate
4290 some large reply buffers so that these replies can be sent even
4291 under memory pressure.
4295 Frequency : rare (liblustre)
4296 Description: fsx running with liblustre complained that using truncate() to
4297 extend the file doesn't work. This patch corrects that issue.
4298 Details : This is the liblustre equivalent of the fix for bug 6196. Fixes
4299 ATTR_SIZE and lsm use in llu_setattr_raw.
4303 Frequency : rare, only 2.6 kernels
4304 Description: Unusual file access patterns on the MDS may result in inode
4305 data being lost in very rare circumstances.
4306 Details : Bad interaction between the ea-in-inode patch and the "no-read"
4307 code in the 2.6 kernel caused the inode and/or EA data not to
4308 be read from disk, causing single-file corruption.
4312 Frequency : rare, only 2.6 filesystems using extents
4313 Description: Heavy concurrent write and delete load may cause data corruption.
4314 Details : It was possible under high-load situations to have an extent
4315 metadata block in the block device cache from a just-unlinked
4316 file overwrite a newly-allocated data block. We now unmap any
4317 metadata buffers that alias just-allocated data blocks.
4321 Frequency : filesystems with default stripe_count larger than 77
4322 Description: lconf+mke2fs fail when formatting filesystem with > 77 stripes
4323 Details : lconf specifies an inode size of 4096 bytes when the default
4324 stripe_count is larger than 77. This conflicts with the default
4325 inode density of 1 per 4096 bytes. Allocate smaller inodes in
4326 this case to avoid pinning too much memory for large EAs.
4328 ------------------------------------------------------------------------------
4330 2005-07-07 Cluster File Systems, Inc. <info@clusterfs.com>
4335 Frequency : rare (extremely heavy IO load with hundreds of clients)
4337 Description: Client is evicted, gets IO error writing to file
4338 Details : lock ordering changes for bug 5492 reintroduced bug 3267 and
4339 caused clients to be evicted for AST timeouts. The fixes in
4340 bug 5192 mean we no longer need to have such short AST timeouts
4341 so ldlm_timeout has been increased.
4344 Frequency : occasional during --force or --failover shutdown under load
4345 Bugzilla : 5949, 4834
4346 Description: Server oops/LBUG if stopped with --force or --failover under load
4347 Details : a collection of import/export refcount and cleanup ordering
4348 issues fixed for safer force cleanup
4351 Frequency : only filesystems larger than 120 OSTs
4352 Bugzilla : 5990, 6223
4353 Description: lfs getstripe would oops on a very large filesystem
4354 Details : lov_getconfig used kfree on vmalloc'd memory
4357 Frequency : only filesystems exporting via NFS to Solaris 10 clients
4358 Bugzilla : 6242, 6243
4359 Description: reading from files that had been truncated to a non-zero size
4360 but never opened returned no data
4361 Details : ll_file_read() reads zeros from no-object files to EOF
4366 Description: A bug in MDS/OSS recovery could cause the OSS to fail an assertion
4367 Details : There's little harm in aborting MDS/OSS recovery and letting it
4368 try again, so I removed the LASSERT and return an error instead.
4370 Severity : enhancement
4372 Description: New debugging infrastructure for tracking down data corruption
4373 Details : The I/O checksum code was replaced to: (a) control it at runtime,
4374 (b) cover more of the client-side code path, and (c) try to narrow
4375 down where problems occurred
4379 Bugzilla : 3819, 4364, 4397, 6313
4380 Description: Racing close and eviction MDS could cause assertion in mds_close
4381 Details : It was possible to get multiple mfd references during close and
4382 client eviction, leading to one thread referencing a freed mfd.
4384 Severity: : enhancement
4385 Bugzilla : 3262, 6359
4386 Description: Attempts to reconnect to servers are now more aggressive.
4387 Details : This builds on the enhanced upcall-less recovery that was added
4388 in 1.4.2. When trying to reconnect to servers, clients will
4389 now try each server in the failover group every 10 seconds. By
4390 default, clients would previously try one server every 25 seconds.
4395 Description: After recovery, certain operations trigger a failed
4396 assertion on a client.
4397 Details : Failing over an mds, using lconf -d --failover, while a
4398 client was doing a readdir() call would cause the client to
4399 LBUG after recovery completed and the readdir() was resent.
4401 Severity : enhancement
4403 Description: Default groups are now added by lconf
4404 Details : You can now run lconf --group <servicename> without having to
4405 manually add groups with lmc.
4408 Frequency : occasional
4410 Description: Nodes with an elan id of 0 trigger a failed assertion
4413 Frequency : always when accessing e.g. tty/console device nodes
4415 Description: tty and some other devices nodes cannot be used on lustre
4416 Details : file's private_data field is used by device data and lustre
4417 values in there got lost. New field was added to struct file to
4418 store fs-specific private data.
4421 Frequency : when exporting Lustre via NFS
4423 Description: NFSD failed occasionally when looking up a path component
4424 Details : NFSD is looking up ".." which was broken in ext3 directories
4425 that had grown large enough to become hashed.
4428 Frequency : Clusters with multiple interfaces not on the same subnet
4430 Description: Nodes will repeatedly try to reconnect to an interface which it
4431 cannot reach and report an error to the log.
4432 Details : Extra peer list entries will be created by lconf with some peers
4433 unreachable. lconf now validates the peer before adding it.
4436 Frequency : Only if a default stripe is set on the filesystem root.
4438 Description: Setting a default stripe on the filesystem root prevented the
4439 filesystem from being remounted.
4440 Details : The client was sending extra request flags in the root getattr
4441 request and did not allocate a reply buffer for the dir EA.
4444 Frequency : occasional, higher if lots of files are accessed by one client
4445 Bugzilla : 6159, 6097
4446 Description: Client trips assertion regarding lsm mismatch/magic
4447 Details : While revalidating inodes the VFS looks up inodes with ifind()
4448 and in rare cases can find an inode that is being freed.
4449 The ll_test_inode() code will free the lsm during ifind()
4450 when it finds an existing inode and then the VFS later attaches
4451 this free lsm to a new inode.
4455 Bugzilla : 6422, 7030
4456 Description: MDS deadlock between mkdir and client eviction
4457 Details : Creating a new file via mkdir or mknod (starting a transaction
4458 and getting the ns lock) can deadlock with client eviction
4459 (gets ns lock and trying to finish a synchronous transaction).
4462 Frequency : occasional
4463 Description: While starting a server, the fsfilt_ext3 module could not be
4465 Details : CFS's improved ext3 filesystem is named ldiskfs for 2.6
4466 kernels. Previously, lconf would still use the ext3 name
4467 when trying to load modules. Now, it will correctly use
4468 ext3 on 2.4 and ldiskfs on 2.6.
4470 Severity : enhancement
4471 Description: The default stripe count has been changed to 1
4472 Details : The interpretation of the default stripe count (0, to lfs
4473 or lmc) has been changed to mean striping across a single
4474 OST, rather than all available. For general usage we have
4475 found a stripe count of 1 or 2 works best.
4477 Severity : enhancement
4478 Description: Add support for compiling against Cray portals.
4479 Details : Conditional compiling for some areas that are different
4483 Frequency : occasional
4484 Bugzilla : 6409, 6834
4485 Description: Creating files with an explicit stripe count may lead to
4486 a failed assertion on the MDS
4487 Details : If some OSTs are full or unavailable, creating files may
4488 trigger a failed assertion on the MDS. Now, Lustre will
4489 try to use other servers or return an error to the
4493 Frequency : occasional
4495 Description: Multiple concurrent overlapping read+write on multiple SMP nodes
4496 caused lock timeout during readahead (since 1.4.2).
4497 Details : Processes doing readahead might match a lock that hasn't been
4498 granted yet if there are overlapping and conflicting lock
4499 requests. The readahead process waits on ungranted lock
4500 (original lock is CBPENDING), while OST waits for that process
4501 to cancel CBPENDING read lock and eventually evicts client.
4503 Severity : enhancement
4505 Description: Initial enabling of flock support for clients
4506 Details : Implements fcntl advisory locking and file status functions.
4507 This feature is provided as an optional mount flag (default
4508 off), and is NOT CURRENTLY SUPPORTED. Not all types of record
4509 locking are implemented yet, and those that are are not guaranteed
4510 to be completely correct in production environments.
4511 mount -t lustre -o [flock|noflock] ...
4514 Frequency : occasional
4516 Description: OSTs running 2.4 kernels but with extents enabled might trip an
4517 assertion in the ext3 JBD (journaling) layer.
4518 Details : The b_committed_data struct is protected by the big kernel lock
4519 in 2.4 kernels, serializing journal_commit_transaction() and
4520 ext3_get_block_handle->ext3_new_block->find_next_usable_block()
4521 access to this struct. In 2.6 kernels there is finer grained
4522 locking to improve SMP performance of the JBD layer.
4526 Description: Changes the "SCSI I/O Stats" kernel patch to default to "enabled"
4531 Description: merge and cleanup kernel patches.
4534 -----------------------------------------------------------------------------
4536 2005-05-05 Cluster File Systems, Inc. <info@clusterfs.com>
4538 NOTE: Lustre 1.4.2 uses an incompatible network protocol than previous
4539 versions of Lustre. Please update all servers and clients to
4540 version 1.4.2 or later at the same time. You must also run
4541 "lconf --write-conf {config}.xml" on the MDS while it is stopped
4542 to update the configuration logs.
4544 - fix for HPUX NFS client breakage when NFS exporting Lustre (5781)
4545 - mdc_enqueue does not need max_mds_easize request buffer on send (5707)
4546 - swab llog records of type '0' so we get proper header size/idx (5861)
4547 - send llog cancel req to DLM cancel portal instead of cb portal (5515)
4548 - fix rename of one directory over another leaking an inode (5953)
4549 - avoid SetPageDirty on 2.6 (5981)
4550 - don't re-add just-being-destroyed locks to the waiting list (5653)
4551 - when creating new directories, inherit the parent's custom
4552 striping settings if present parent (3048)
4553 - flush buffers from cache before direct IO in 2.6 obdfilter (4982)
4554 - don't hold i_size_sem in ll_nopage() and ll_ap_refresh_count (6077)
4555 - don't hold client locks on temporary worklist from l_lru (5666)
4556 - handle IO errors in 2.6 obdfilter bio completion routine (6046)
4557 - automatically evict dead clients (5921)
4558 - Update file size properly in create+truncate+fstat case (6196)
4559 - Do not unhash mountpoint dentries, do not allow removal of
4561 - Avoid lock ordering deadlock issue with write/truncate (6203,5654)
4562 - reserve enough journal credits in fsfilt_start_log for setattr (4554)
4563 - ldlm_enqueue freed-export error path would always LBUG (6149,6184)
4564 - don't reference lr_lvb_data until after we hold lr_lvb_sem (6170)
4565 - don't overwrite last_rcvd if there is a *_client_add() error (6086)
4566 - Correctly handle reads of files with no objects (6243)
4567 - lctl recover will also mark a device active if deactivate used (5933)
4569 - by default create 1 inode per 4kB space on MDS, per 16kB on OSTs
4570 - allow --write-conf on an MDS with different nettype than client (5619)
4571 - don't write config llogs to MDS for mounts not from that MDS (5617)
4572 - lconf should create multiple TCP connections from a client (5201)
4573 - init scripts are now turned off by default; run chkconfig --on
4574 lustre and chkconfig --on lustrefs to use them
4575 - upcalls are no longer needed for clients to recover to failover
4577 - add --abort-recovery option to lconf to abort recovery on device
4579 - add support for an arbitrary number of OSTs (3026)
4580 - Quota support protocol changes.
4581 - forward compatibility changes to wire structs (6007)
4582 - rmmod NALs that might be loaded because of /etc/modules.conf (6133)
4583 - support for mountfsoptions and clientoptions to the Lustre LDAP (5873)
4584 - improved "lustre status" script
4585 - initialize blocksize for non-regular files (6062)
4586 - added --disable-server and --disable-client configure options (5782)
4587 - introduce a lookup cache for lconf to avoid repeated DB scans (6204)
4588 - Vanilla 2.4.29 support
4589 - increase maximum number of obd devices to 520 (6242)
4590 - remove the tcp-zero-copy patch from the suse-2.4 series (5902)
4591 - Quadrics Elan drivers are now included for the RHEL 3 2.4.21 and
4592 SLES 9 2.6.5 kernels
4593 - limit stripes per file to 160 (the maximum EA size) (6093)
4595 2005-03-22 Cluster File Systems, Inc. <info@clusterfs.com>
4598 - don't LASSERT in ll_release on NULL lld with NFS export (4655, 5760)
4599 - hold NS lock when calling handle_ast_error->del_waiting_lock (5746)
4600 - fix setattr mtime regression from lovcleanup merge (4829, 5669)
4601 - workaround for 2.6 crash in ll_unhash_aliases (5687, 5210)
4602 - small ext3 extents cleanups and fixes (5733)
4603 - improved mballoc code, several small races and bugs fixed (5733, 5638)
4604 - kernel version 43 - fix remove_suid bugs in both 2.4 and 2.6 (5695)
4605 - avoid needless client->OST connect, fix handle mismatch (5317)
4606 - fix DLM error path that led to out-of-sync client, long delays (5779)
4607 - support common vfs-enforced mount options (nodev,nosuid,noexec) (5637)
4608 - fix several locking issues related to i_size (5492,5624,5654,5672)
4609 - don't move pending lock onto export if it is already evicted (5683)
4610 - fix kernel oops when creating .foo in unlinked directory (5548)
4611 - fix deadlock in obdfilter statistics vs. object create (5811)
4612 - use time_{before,after} to avoid timer jiffies wrap (5882)
4613 - shutdown --force/--failover stability (3607,3651,4797,5203,4834)
4614 - Do not leak request if server was not able to process it (5154)
4615 - If mds_open unable to find parent dir, make that negative lookup(5154)
4616 - don't create new directories with extent-mapping (5909, 5936)
4618 - fix lustre/lustrefs init scripts for SuSE (patch from Scali, 5702)
4619 - don't hold the pinger_sem in ptlrpc_pinger_sending_on_import
4620 - change obd_increase_kms to obd_adjust_kms (up or down) (5654)
4621 - lconf, lmc search both /usr/lib and /usr/lib64 for Python libs (5800)
4622 - support for RHEL4 kernel on i686 (5773)
4623 - provide error messages when incompatible logs are encountered (5898)
4625 2005-02-18 Cluster File Systems, Inc. <info@clusterfs.com>
4626 * version 1.4.0.10 (1.4.1 release candidate 1)
4628 - don't keep a lock reference when lock is not granted (4238)
4629 - unsafe list practices (rarely) led to infinite eviction loop (4908)
4630 - add per-fs limit of Lustre pages in page cache, avoid OOM (4699)
4631 - drop import inflight refcount on signal_completed_replay error (5255)
4632 - unlock page after async write error during send (3677)
4633 - handle missing objects in filter_preprw_read properly (5265)
4634 - no transno return for symlink open, don't save no-trasno open (3440)
4635 - don't try to complete elan receive that already failed (4012)
4636 - free RPC server reply state on error (5406)
4637 - clean up thread from ptlrpc_start_thread() on error (5160)
4638 - readahead could read extra page into cache that wasn't ejected (5388)
4639 - prevent races in class_attach/setup/cleanup/detach (5260)
4640 - don't dereference de->d_inode after l_dput of de (5458)
4641 - use "int" for stripe value returned from lock_to_stripe (5544)
4642 - mballoc allocation and error-checking fixes in 2.6 (5504)
4643 - block device patches to fix I/O request sizes in 2.6 (5482)
4644 - look up hostnames for IB nals (5602)
4645 - 2.6 changed lock ordering of 2 semaphores, caused deadlock (5654)
4646 - don't start multiple acceptors for the same port (5277)
4647 - fix incorrect LASSERT in mds_getattr_name (5635)
4648 - export a proc file for general "ping" checking (5628)
4649 - fix "lfs check" to not block when the MDS is down (5628)
4651 - service request history (4965)
4652 - put {ll,lov,osc}_async_page structs in a single slab (4699)
4653 - create an "evict_client" /proc entry on OSTs, like the MDS has
4654 - fix mount usage message, return errors per mount(8) (5168)
4655 - change grep [] to grep "[]" in tests so they work in more UMLs
4656 - fix ppc64/x86_64 spec to use %{_libdir} instead of /usr/lib (5389)
4657 - remove ancient LOV_MAGIC_V0 EA support (5047)
4658 - add "disk I/Os in flight" and "I/O req time" stats in obdfilter
4659 - align r/w RPCs to PTLRPC_MAX_BRW_SIZE boundary for performance (3451)
4660 - allow readahead allocations to fail when low on memory (5383)
4661 - mmap locking landed again, after considerable improvement (2828)
4662 - add get_hostaddr() to lustreDB.py for LDAP support (5459)
4664 2004-11-23 Cluster File Systems, Inc. <info@clusterfs.com>
4667 - send OST transaction number in read/write reply to free req (4966)
4668 - don't ASSERT in ptl_send_rpc() if we run out of memory (5119)
4669 - lock /proc/sys/portals/routes internal state, avoiding oops (4827)
4670 - the watchdog thread now runs as interruptible (5246)
4671 - flock/lockf fixes (but it's still disabled, pending 5135)
4672 - don't use EXT3 constants in llite code (5094)
4673 - memory shortage at startup could cause assertion (5176)
4675 - reorganization of lov code
4676 - single portals codebase
4678 - add extents/mballoc support (5025)
4679 - direct I/O reads in the obdfilter (4048)
4680 - kernel patches from LNXI for 2.6 (bluesmoke, perfctr, mtd, kexec)
4682 tbd Cluster File Systems, Inc. <info@clusterfs.com>
4685 - send OST transaction number in read/write reply to free req (4966)
4686 - don't ASSERT in ptl_send_rpc() if we run out of memory (5119)
4687 - lock /proc/sys/portals/routes internal state, avoiding oops (4827)
4688 - the watchdog thread now runs as interruptible (5246)
4689 - handle missing objects in filter_preprw_read properly (5265)
4690 - unsafe list practices (rarely) led to infinite eviction loop (4908)
4691 - drop import inflight refcount on signal_completed_replay error (5255)
4692 - unlock page after async write error during send (3677)
4693 - return original error code on reconstructed replies (3761)
4694 - no transno return for symlink open, don't save no-trasno open (3440)
4696 - add pid to ldlm debugging output (4922)
4697 - bump the watchdog timeouts -- we can't handle 30sec yet
4698 - extra debugging for orphan dentry/inode bug (5259)
4700 2004-11-16 Cluster File Systems, Inc. <info@clusterfs.com>
4703 - fix TCP_NODELAY bug, which caused extreme perf regression (5134)
4704 - allocate qswnal tx descriptors singly to avoid fragmentation (4504)
4705 - don't LBUG on obdo_alloc() failure, use OBD_SLAB_ALLOC() (4800)
4706 - fix NULL dereference in /proc/sys/portals/routes (4827)
4707 - allow failed mdc_close() operations to be interrupted (4561)
4708 - stop precreate on OST before MDS would time out on it (4778)
4709 - don't send partial-page writes before EOF from client (4410)
4710 - discard client grant for sub-page writes on large-page clients (4520)
4711 - don't free dentries not owned by NFS code, check generation (4806)
4712 - fix lsm leak if mds_create_objects() fails (4801)
4713 - limit debug_daemon file size, always print CERROR messages (4789)
4714 - use transno after validating reply (3892)
4715 - process timed out requests if import state changes (3754)
4716 - update mtime on OST during writes, return in glimpse (4829)
4717 - add mkfsoptions to LDAP (4679)
4718 - use ->max_readahead method instead of zapping global ra (5039)
4719 - don't interrupt __l_wait_event() during strace
4721 - add software watchdogs to catch hung threads quickly (4941)
4722 - make lustrefs init script start after nfs is mounted
4723 - fix CWARN/ERROR duplication (4930)
4724 - return async write errors to application if possible (2248)
4725 - add /proc/sys/portal/memused (bytes allocated by PORTALS_ALLOC)
4726 - print NAL number in %x format (4645)
4727 - update barely-supported suse-2.4.21-171 series (4842)
4728 - support for sles 9 %post scripts
4729 - support for building 2.6 kernel-source packages
4730 - support for sles km_* packages
4732 2004-10-07 Cluster File Systems, Inc. <info@clusterfs.com>
4735 - ignore -ENOENT errors in osc_destroy (3639)
4736 - notify osc create thread that OSC is being cleaned up (4600)
4737 - add nettype argument for llmount in #5d in conf-sanity.sh (3936)
4738 - reconstruct ost_handle() like mds_handle() (4657)
4739 - create a new thread to do import eviction to avoid deadlock (3969)
4740 - let lconf resolve symlinked-to devices (4629)
4741 - don't unlink "objects" from directory with default EA (4554)
4742 - hold socknal file ref over connect in case target is down (4394)
4743 - allow more than 32000 subdirectories in a single directory (3244)
4744 - fix blocks count for O_DIRECT writes (3751)
4745 - OST returns ENOSPC from object create when no space left (4539)
4746 - don't send truncate RPC if file size isn't changing (4410)
4747 - limit OSC precreate to 1/2 of value OST considers bogus (4778)
4748 - bind to privileged port in socknal and tcpnal (3689)
4750 - rate limit CERROR/CWARN console message to avoid overload (4519)
4751 - GETFILEINFO dir ioctl returns LOV EA + MDS stat in 1 call (3327)
4752 - basic mmap support (3918)
4753 - kernel patch series update from b1_4 (4711)
4755 2004-09-16 Cluster File Systems, Inc. <info@clusterfs.com>
4758 - avoid crash during MDS cleanup with OST shut down (2775)
4759 - fix loi_list_lock/oig_lock inversion on interrupted IO (4136)
4760 - don't use bad inodes on the MDS (3744)
4761 - dynamic object preallocation to improve recovery speed (4236)
4762 - don't hold spinlock over lock dumping or change debug flags (4401)
4763 - don't zero obd_dev when it is force cleaned (3651)
4764 - print grants to console if they go negative (4431)
4765 - "lctl deactivate" will stop automatic recovery attempts (3406)
4766 - look for existing locks in ldlm_handle_enqueue() (3764)
4767 - don't resolve lock handle twice in recovery avoiding race (4401)
4768 - revalidate should check working dir is a directory (4134)
4770 - don't always mark "slow" obdfilter messages as errors (4418)
4772 2004-08-24 Cluster File Systems, Inc. <info@clusterfs.com>
4775 - don't close LustreDB during write_conf until it is done (3860)
4776 - fix typo in lconf for_each_profile (3821)
4777 - allow dumping logs from multiple threads at one time (3820)
4778 - don't allow multiple threads in OSC recovery (3812)
4779 - fix debug_size parameters (3864)
4780 - fix mds_postrecov to initialize import for llog ctxt (3121)
4781 - replace config semaphore with spinlock (3306)
4782 - be sure to send a reply for a CANCEL rpc with bad export (3863)
4783 - don't allow enqueue to complete on a destroyed export (3822)
4784 - down write_lock before checking llog header bitmap (3825)
4785 - recover from lock replay timeout (3764)
4786 - up llog sem before sending rpc (3652)
4787 - reduce ns lock hold times when setting kms (3267)
4788 - change a dlm LBUG to LASSERTF, to maybe learn something (4228)
4789 - fix NULL deref and obd_dev leak on setup error (3312)
4790 - replace some LBUG about llog ops with error handling (3841)
4791 - don't match INVALID dentries from d_lookup and spin (3784)
4792 - hold dcache_lock while marking dentries INVALID and hashing (4255)
4793 - fix invalid assertion in ptlrpc_set_wait (3880)
4795 - add libwrap support for the TCP acceptor (3996)
4796 - add /proc/sys/portals/routes for non-root route listing (3994)
4797 - allow setting MDS UUID in .xml (2580)
4798 - print the stack of a process that LBUGs (4228)
4800 2004-07-14 Cluster File Systems, Inc. <info@clusterfs.com>
4803 - don't cleanup request in ll_file_open() on failed MDS open (3430)
4804 - make sure to unset replay flag from failed open requests (3440)
4805 - if default stripe count is 0, use OST count for inode size (3636)
4806 - update parent mtime/ctime on client for create/unlink (2611)
4807 - drop dentry ref in ext3_add_link from open_connect_dentry (3266)
4808 - free recovery state on server during a forced cleanup (3571)
4809 - unregister_reply for resent reqs (3063)
4810 - loop back devices mounting and status check on 2.6 (3563)
4811 - fix resource-creation race that can provoke i_size == 0 (3513)
4812 - don't try to use bad inodes returned from MDS/OST fs lookup (3688)
4813 - more debugging for page-accounting assertion (3746)
4814 - return -ENOENT instead of asserting if ost getattr+unlink race (3558)
4815 - avoid deadlock after precreation failure (3758)
4816 - fix race and lock order deadlock in orphan handling (3450, 3750)
4817 - add validity checks when grabbing inodes from l_ast_data (3599)
4819 - add /proc/.../recovery_status to obdfilter (3428)
4820 - lightweight CDEBUG infrastructure, debug daemon (3668)
4821 - change default OSC RPC parameters to be better on small clusters
4822 - turn off OST read cache for files smaller than 32MB
4823 - install man pages and include them in rpms (3100)
4824 - add new init script for (un)mounting lustre filesystems (2593)
4825 - run chkconfig in %post for init scripts (3701)
4826 - drop scimac NAL (unmaintained)
4828 2004-06-17 Cluster File Systems, Inc. <info@clusterfs.com>
4831 - clean kiobufs before and after use (3485)
4832 - strip trailing '/'s before comparing paths with /proc/mounts (3486)
4833 - remove assertions to work around "in-flight rpcs" recovery bug (3063)
4834 - change init script to fail more clearly if not run as root (1528)
4835 - allow clients to reconnect during replay (1742)
4836 - fix ns_lock/i_sem lock ordering deadlock for kms update (3477)
4837 - don't do DNS lookups on NIDs too small for IP addresses (3442)
4838 - re-awaken ptlrpcd if new requests arrive during check_set (3554)
4839 - fix cond_resched (3554)
4840 - only evict unfinished clients after recovery (3515)
4841 - allow bulk resend, prevent data loss (3570)
4842 - dynamic ptlrpc request buffer allocation (2102)
4843 - don't allow unlinking open directory if it isn't empty (2904)
4844 - set MDS/OST threads to umask 0 to not clobber client modes (3359)
4845 - remove extraneous obd dereference causing LASSERT failure (3334)
4846 - don't use get_cycles() when creating temp. files on the mds (3156)
4847 - hold i_sem when setting i_size in ll_extent_lock() (3564)
4848 - handle EEXIST for set-stripe, set proper directory name (3336)
4850 - servers can dump a log evicting a client - lustre.dump_on_timeout=1
4851 - fix ksocknal_fmb_callback() error messages (2918)
4853 2004-05-27 Cluster File Systems, Inc. <info@clusterfs.com>
4856 - don't copy lvb into (possibly NULL) reply on error (2983)
4857 - don't deref dentry after dput, don't free lvb on error (2922)
4858 - use the kms to determine writeback rpc length (2947)
4859 - increment oti_logcookies when osc is inactive (2948)
4860 - update client's i_blocks count via lvb messages (2543)
4861 - handle intent open/close of special files properly (1557)
4862 - mount MDS with errors=remount-ro, like obdfilter (2009)
4863 - initialize lock handle to avoid ASSERT on error cleanup (3057)
4864 - don't use cancelling-locks' kms values (2947)
4865 - use highest lock extent for kms, not last one (2925)
4866 - don't dereference ERR_PTR() dentry in error handling path (3107)
4867 - fix thread race in portals_debug_dumplog() (3122)
4868 - create lprocfs device entries at setup instead of at attach (1519)
4869 - common AST error handler, don't evict client on completion race (3145)
4870 - zero nameidata in detach_mnt in 2.6 (3118)
4871 - verify d_inode after revalidate_special is valid in 2.6 (3116)
4872 - use lustre_put_super() to handle zconf unmounts in 2.6 (3064)
4873 - initialize RPC timeout timer earlier for 2.6 (3219)
4874 - don't dereference NULL reply buffer if mdc_close was never sent (2410)
4875 - print nal/nid for unknown nid (3258)
4876 - additional checks for oscc recovery before doing precreate (3284)
4877 - fix ll_extent_lock() error return code for 64-bit systems (3043)
4878 - don't crash in mdc_close for bad permissions on open (3285)
4879 - zero i_rdev for non-device files (3147)
4880 - clear page->private before handing to FS, better assertion (3119)
4881 - tune the read pipeline (3236)
4882 - fix incorrect decref of invalidated dentry (2350)
4883 - provide read-ahead stats and refine rpc in flight stats (3328)
4884 - don't hold journal transaction open across create RPC (3313)
4885 - update atime on MDS at close time (3265)
4886 - close LDAP connection when recovering to avoid server load (3315)
4887 - update iopen-2.6 patch with fixes from 2399,2517,2904 (3301)
4888 - don't leak open file on MDS after open resend (3325)
4889 - serialize filter_precreate and filter_destroy_precreated (3329)
4890 - loop device shouldn't call sync_dev() for nul device (3092)
4891 - clear page cache after eviction (2766)
4892 - resynchronize MDS->OST in background (2824)
4893 - refuse to mount the same filesystem twice on same mountpoint (3394)
4894 - allow llmount to create routes for mounting behind routers (3320)
4895 - push lock cancellation to blocking thread for glimpse ASTs (3409)
4896 - don't call osc_set_data_with_check() for TEST_LOCK matches (3159)
4897 - fix rare problem with rename on htree directories (3417)
4899 - allow default OST striping configuration per directory (1414)
4900 - fix compilation for qswnal for 2.6 kernels (3125)
4901 - increase maximum number of MDS request buffers for large systems
4902 - change liblustreapi to be useful for external progs like lfsck (3098)
4903 - increase local configuration timeout for slow disks (3353)
4904 - allow configuring ldlm AST timeout - lustre.ldlm_timeout=<seconds>
4906 2004-03-22 Cluster File Systems, Inc. <info@clusterfs.com>
4909 - fixes for glimpse AST timeouts / incorrectly 0-sized files (2818)
4910 - don't overwrite extent policy data in reply if lock was blocked (2901)
4911 - drop filter export grants atomically with removal from device (2663)
4912 - del obd_self_export from work_list in class_disconnect_exports (2908)
4913 - don't LBUG if MDS recovery times out during orphan cleanup (2530)
4914 - swab reply message in mdc_close, other PPC fixes (2464)
4915 - fix destroying of named logs (2325)
4916 - overwrite old logs when running lconf --write_conf (2264)
4917 - bump LLOG_CHUNKSIZE to 8k to allow for larger clusters (2306)
4918 - fix race in target_handle_connect (2898)
4919 - mds_reint_create() should take same inode create lock (2926)
4920 - correct journal credits calculated for CANCEL_UNLINK_LOG (2931)
4921 - don't close files for self_export to avoid uninitialized obd (2936)
4922 - allow MDS with the same name as client node (2939)
4923 - hold dentry reference for closed log files for unlink (2325)
4924 - reserve space for all logs during transactions (2059)
4925 - don't evict page beyond end of stripe extent (2925)
4926 - don't oops on a deleted current working directory (2399)
4927 - handle hard links to targets without a parent properly (2517)
4928 - don't dereference NULL lock when racing during eviction (2867)
4929 - don't grow lock extents when lots of conflicting locks (2919)
4931 2004-03-04 Cluster File Systems, Inc. <info@clusterfs.com>
4934 - account for cache space usage on clients to avoid data loss (974)
4935 - lfsck support in lustre kernel code (2349)
4936 - reduce journal credits needed for BRW writes (2370)
4937 - orphan handling to avoid losing space on client/server crashes
4938 - ptlrpcd can be blocked, stopping ALL progress (2477)
4939 - use lock value blocks to assist in proper KMS, faster stat (1021)
4940 - takes i_sem instead of DLM locks internally on obdfilter (2720)
4941 - recovery for initial connections (2355)
4942 - fixes for mds_cleanup_orphans (1934)
4943 - abort_recovery crashes MDS in b_eq (mds_unlink_orphan) (2584)
4944 - block all file creations until orphan recovery completes (1901)
4945 - client remove rq_connection from request struct (2423)
4946 - conf-sanity test_5, proper cleanup in umount log not availale (2640)
4947 - recovery timer race (2670)
4948 - mdc_close recovey bug (2532)
4949 - ptlrpc cleanup bug (2710)
4950 - mds timeout on local locks (2588)
4951 - namespace lock held during RPCs (2431)
4952 - handle interrupted sync write properly (2503)
4953 - don't try to handle a message that hasn't been replied to (2699)
4954 - client assert failure during cleanup after abort recovery (2701)
4955 - leak mdc device after failed mount (2712)
4956 - ptlrpc_check_set allows timedout requests to complete (2714)
4957 - wait for inflight reqs when ptlrpcd finishes (2710)
4958 - make sure unregistered services are removed from the srv_list
4959 - reset bulk XID's when resending them (caught by 1138 test)
4960 - unregister_bulk after timeout
4961 - fix lconf error (2694)
4962 - handle write after unfinished setstripe, stripe-only getstripe (2388)
4963 - readahead locks pages, leaves pending causing memory pressure (2673)
4964 - increase OST request buffers to 4096 on large machines (2729)
4965 - fix up permission of existing directories in simple_mkdir (2661)
4966 - init deleted item, add assertions ptlrpc_abort_inflight() (2725)
4967 - don't assign transno to errored transactions (2742)
4968 - don't delete objects on OST if given a bogus objid from MDS (2751)
4969 - handle large client PAGE_SIZE readdir on small PAGE_SIZE MDS (2777)
4970 - if rq_no_resend, then timeout request after recovery (2432)
4971 - fix MDS llog_logid record size, 64-bit array alignment (2733)
4972 - don't call usermode_helper from ptlrpcd, DEFAULT upcall (2773)
4973 - put magic in mount.lustre data, check for bad/NULL mount data (2529)
4974 - MDS recovery shouldn't delete objects that it has given out (2730)
4975 - if enqueue arrives after completion, don't clobber LVB (2819)
4976 - don't unlock pages twice when trigger_group_io returns error (2814)
4977 - don't deref NULL rq_repmsg if ldlm_handle_enqueue failed (2822)
4978 - don't write pages to disk if there was an error (1450)
4979 - don't ping imports that have recovery disabled (2676)
4980 - take buffered bytes into account when balancing socknal conn (2817)
4981 - hold a DLM lock over readdir always, use truncate_inode_pages (2706)
4982 - reconnect unlink llog connection after MDS reconnects to OST (2816)
4983 - remove little-endian swabbing of llog records (1987)
4984 - set/limit i_blksize to LL_MAX_BLKSIZE on client (2884)
4985 - retry reposting request buffers if they fail (1191)
4986 - grow extent at grant time to avoid granting a revoked lock (2809)
4987 - lock revoke doesn't evict page if covered by a second lock (2765)
4988 - disable VM readahead to avoid reading outside lock extents (2805)
4990 - return LL_SUPER_MAGIC from statfs for the filesystem type (1972)
4991 - updated kernel patches for hp-2.4.20 kernel (2681)
4993 2004-02-07 Cluster File Systems, Inc. <info@clusterfs.com>
4996 - fix truncated write corruption (2366)
4997 - fix for failed assertion in iopen_connect_dentry (1792,2517)
4999 - don't flag the ptlrpcd thread with PF_MEMALLOC (2636)
5000 - ensure len(uuid) < 37 in lmc (1171)
5001 - fix ia64 OOPS in llog_test (2255)
5002 - zero end of page at obdfilter for partial page writes (2648)
5003 - don't leave stale dentries around after renames (bug 2428)
5004 - fix timeouts when evicting a client with a single lock held (2642)
5005 - set deadline for the initial HELLO message to drain (2634)
5006 - print out dotted-quad IP addresses in the socknal (2302)
5008 - additional debugging for MDS client eviction problem (2443)
5009 - fix mkfsoptions support for osts (2603, 2604)
5011 2004-01-27 Cluster File Systems, Inc. <info@clusterfs.com>
5014 - add series for the vanilla 2.6.0 kernel
5015 - add series for the vanilla 2.4.24 kernel
5016 - add series for a cray x86/64 UL kernel drop
5017 - fix xattr patches for the vanilla 2.4.19 series
5019 - generate true UUIDs in lmc (1171)
5020 - have portals stack dumping break in UML (2466)
5021 - avoid bad dchild deref; avoid inum lock w/o creation (2362)
5022 - allocate with _NOFS in ldlm to avoid deadlock (1933)
5023 - wake callback waiting threads on client eviction (2460)
5024 - Add --ptldebug and --subsystem to lmc (1719)
5025 - update assertion to allow safe interrupt allocation
5026 - set rq_no_resend for cancel requests (2432)
5027 - recalculate ptlrpcd timeout after resend (2494)
5028 - call vfs_rmdir when removing pending directories (2368)
5029 - fix renaming a file to itself (2429)
5030 - lmc creates a default one-stripe lov (2454)
5031 - expand procfs space to handle large clusters (2326)
5032 - increase UML stack to avoid overflow
5033 - update lconf's list of debug and subsystem masks
5034 - fix lfs find --obd (2510)
5035 - /proc tunable for disabling filter read caching (2591)
5036 - stop rpm packages from altering slapd.conf (2301)
5037 - disable nagle in the socknal under 0conf (2578)
5038 - choose mds inode size based on stripe count (2572)
5039 - fix kernel-source rpm problems (2516)
5041 - add --disable-doc to avoid pdf generation (2421)
5042 - update documentation, tests, type-os, comments
5043 - avoid format warnings on ia64
5044 - remove the TOE NAL
5045 - tiny code cleanups by removing unused fields
5047 2004-01-07 Cluster File Systems, Inc. <info@clusterfs.com>
5050 - fix obvious semaphore misuse in as-yet-unused setattr path (2348)
5051 - remove the most blatant lies from BUILDING file (2371)
5052 - change default debug level to reasonable production setting
5053 - reduce client side cache size to reduce cache flush time
5054 - reduce max RPCs in flight to avoid unnecessary file fragmentation
5055 - make TCP zerocopy and pinger support enabled by default (2476)
5056 - sync writes completed after process exits caused crashes (2319)
5057 - maintain correct mount count on the MDS (2356)
5058 - backout 1557, because 2316 wasn't really fixed
5059 - better file I/O statistics gathering in /proc
5060 - don't take unnecessary, deadlock-inducing bug in readpage (2383)
5061 - another kernel patch to fix zero-copy TCP function export
5062 - don't take duplicate lock when processing re-sent getattr (2420)
5063 - lctl uses obd_self_export instead of creating new conn (2353)
5064 - MDS/OST recovery case which requires object creation asserted (2425)
5065 - move lfs from /usr/sbin to /usr/bin in packages
5066 - fix race between mds_client_add and mds_client_free (2417)
5067 - use kmalloc instead of slabs in portals (2430)
5068 - don't create duplicate records when a failover MDS is present (2442)
5069 - remove unnecessary mount age check (2332)
5070 - don't remove directory inodes from locks prematurely (2451)
5071 - don't break if MDS service name is the same as hostname (2103)
5072 - fix races in client write RPC generation when cache full (2482)
5074 2003-12-13 Cluster File Systems, Inc. <info@clusterfs.com>
5077 - remove now-unused request->rq_obd (278)
5078 - if an allocation fails, print out how much memory we've used (1933)
5079 - use PORTAL_SLAB_ALLOC for structures, to get GFP_MEMALLOC (1933)
5080 - add the "configurable stack size" patch to most series files (1256)
5081 - ability to write large log records, for 100+ OST configs (2306)
5082 - fix NULL deref when filter_prep fails (2314)
5083 - fix operator precedence error in filter_sync
5084 - dynamic allocation of socknal TX descriptors (2315)
5085 - fix a missed case in the GFP_MEMALLOC patch, can cause deadlock (2310)
5086 - fix gcc 2.96 compilation problem in xattr kernel patch (2294)
5087 - ensure that CWARN messages in Portals always get to the syslog
5088 - __init/__exit are not for prototype decls (ldlm_init/exit)
5089 - x86-64 compile warning fixes
5090 - fix gateway LMC keyword conflict (2318)
5091 - fix MDS lock inversions in getattr/reint paths (1844)
5092 - fix a rare lock re-ordering bug, which caused deadlock (2322)
5093 - fix i_sem/journal inversion in fsfilt_ext3_write_record (2306)
5094 - DLM race condition prevented some lock evictions (2328)
5095 - ENOMEM detection and retry on socknal sends (2230)
5096 - use GFP_NOFS throughout Lustre, to combat ENOMEM (2230)
5097 - move osc_rpcd into ptlrpc, for use in MDC and others (2329)
5098 - protect MDS inode fsdata with stronger locking; fixes assertion (2313)
5099 - better error messages when a client is rejected during recovery (1505)
5100 - avoid cancelling locks which were never granted, after failure (2330)
5101 - fix i_sem/journal inversion in mds_client_add (2333)
5102 - fix truncate/getattr lock cycle deadlock (2334)
5103 - use rpcd to send close; allows resend after timeout, avoid leak (1897)
5104 - fix two rare exit paths which could leak an l_lock() ref (2321)
5105 - fencepost error in MDS/OST orphan recovery (2226)
5106 - make log record alignment 8 bytes (1988)
5107 - lstripe now fails when requested offset > ost_count (2237)
5108 - ensure that all kernel series have a complete list.h (1607)
5109 - fix crashes in special-file operations (2316)
5110 - lctl create/brw OID mismatch, caused by obsolete filter loop (2339)
5112 - allow configurable automake binary, for testing new versions
5113 - small update to the lfs documentation
5115 2003-12-03 Cluster File Systems, Inc. <info@clusterfs.com>
5117 * fix negative export reference count in fsfilt_sync (2312)
5119 2003-12-01 Cluster File Systems, Inc. <info@clusterfs.com>
5120 * release candidate 0.9.1
5122 - orphans are moved into the PENDING directory for possible recovery
5123 - replayed opens now open by fid for orphan/rename safety (1042)
5124 - last close of an orphan inode generates a transno (683)
5125 - chdir() and mount() now pin the directory entry (1020)
5126 - avoid CERROR in normal ll_setattr_raw() error case (1500)
5127 - discard very old requests without processing them (1502)
5128 - remove some common, well-understood CERRORs (1505)
5129 - require O_DIRECT I/O to be page-sized to workaround IA64 crash (1609)
5130 - clear "grant" flags in OST replies until OST grant code lands (1644)
5131 - fix read performance by not clobbering i_blksize on client (1598)
5132 - fix __ldlm_handle2lock oops by not dereferencing lock after PUT (1625)
5133 - make LRU size a /proc tunable, clears locks when reduced (707)
5134 - fix some lprocfs rot that prevented ptlbd from loading (1732)
5135 - server locks take references on exports now (1558)
5136 - build fixes for 2.4.20-rh trees (1663)
5137 - return an error from lov_create if all OSCs are inactive (1751)
5138 - fix import levels when a reconnect happens without a timeout (1597)
5139 - exit early from mds_open if we get a lookup error (1749)
5140 - partial page read at EOF wouldn't wait for disk before sending (1642)
5141 - avoid NULL deref in obdfilter when reading page past EOF (1592)
5142 - avoid LASSERT in ll_intent_lock if server failed very early (1090)
5143 - fix LBUG in ll_it_open_error with rc = -2 (1861)
5144 - write/truncate lock inversion (1639)
5145 - Don't auto-load obdclass, portals modules during cleanup (1495)
5146 - fix timestamps from jumping to "now" (1763)
5147 - extra journal assertions (1648)
5148 - add an extra multiunlink test (1771)
5149 - fix read_record/write_record API (1776)
5150 - fix leak of offset_extent, possible incorrect i_size later (1772)
5151 - fix lasserts in mis-matched transnos during open-unlink testing (1541)
5152 - Debugging for the kqswnal_get_idle_tx problems (1820)
5153 - Allow recovery to be attempted multiple times (1536)
5154 - Write out MDS last_rcvd file after it is first created (1600)
5155 - Fix tx_descriptor leak in failed transmit situations (1827)
5156 - ext3 journaling fixes for assertion failure after IO error (1871)
5157 - class_export_put() on freed export after completion AST error (1896)
5158 - Fix revalidate looping in VFS (1322)
5159 - Don't access a freed export during MDS_REINT timeout (1521)
5160 - Add open-unlink recovery support on the MDS (1673,1764)
5161 - Return an error if no MDS data was read from last_rcvd (1946)
5162 - Fix for lookup "." or ".." crash on error (1932,1931,1935)
5163 - Don't setup a disk device that doesn't match exported UUID (317)
5164 - Reduce bulk RPC timeout to avoid cascading client/OST failures (1845)
5165 - avoid committing NULL handle in force close
5166 - local.sh is now a one-stripe LOV configuration
5167 - POSIX utime.4 -EPERM on FIFO not owned by user (56)
5168 - fix ext3 htree duplicate directory entry corruption (1516)
5169 - POSIX creat.13, fstat.1, open.18, stat.3 new file atime/mtime (2020)
5170 - update to new LOV EA format (2097)
5171 - interoperability for different PAGE_SIZE/wordsize (686,1821,1343,2042)
5173 2003-06-15 Phil Schwan <phil@clusterfs.com>
5176 - imports and exports cleanup too early, need refcounts (349, 879, 1045)
5177 - per-import/export recovery handling (958, 931, 959)
5178 - multiple last-rcvd slots, for serving multiple FSes (949)
5179 - connections are again shared between multiple imp/exports (963, 964)
5180 - "umount -f" would hang if any requests needed to be sent (393, 978)
5181 - avoid pinning large req buffer by copying for queued messages (989)
5182 - add "uuid" to "lctl device" command to help upcalls (991)
5183 - "open" RPCs with transnos would confuse recovery counters (1037)
5184 - do proper endian conversion of all wire messages (288, 340, 891)
5185 - remove OST bulk get LBUGs, fix ost_brw_write cleanup (1126)
5186 - call waiting locks callback from LDLM recovery thread (1127, 1151)
5187 - fix ptlrpc_connection leak in target_handle_connect (1174)
5188 - fix import refcounting bug in OST and MDS cleanup (1134)
5189 - if an invalid-at-open-time OSC returned before close(), LBUG (1150)
5190 - fix very unlikely obd_types race condition (501)
5191 - remove osc_open hack for echo_client (1187)
5192 - we leaked exports/dlmimps for forcibly disconnected clients (1143)
5193 - a failure in read_inode2 leads to deadlock (1139)
5194 - cancel ack-locks as soon as transaction is committed (1072)
5195 - fix major leaks and crashes in the bulk I/O path (937, 1057)
5196 - make sure to commitrw after any preprw to avoid deadlock (1162)
5197 - failing to execute a file in a lustre FS would lock inode (1203)
5198 - small DEBUG_REQ fix to avoid dereferencing a NULL (1227)
5199 - don't ASSERT while cleaning up an incompletely-setup obd (1248)
5200 - obd_uuid2tgt would walk off the end of the list (1255)
5201 - on IA64 the osc would give portals incorrect bulk size (1258)
5202 - fix debug daemon ioctl interface; allows daemon on ia64 (1274)
5203 - fix lock inversion caused by new llite matching code (1282)
5204 - limit the number of dirty pages on a client to 10MB (1286)
5205 - timed out locks were not being corrected cancelled (1289)
5206 - fix O_DIRECT above 4GB on IA-32 (1292)
5207 * major user-visible changes
5208 - fail out/fail over policy now controlled by the upcall (993)
5210 - add OBD_PING to check server availability and failure (954)
5211 - lustre messages are now sent in sending host order (288, 340, 891)
5212 - add eadatalen to MDS getattr reply (340)
5213 - OST read replies may contain second buffer, with per-page status (593)
5215 2003-03-11 Phil Schwan <phil@clusterfs.com>
5218 - LDLM_DEBUG macro fix, for gcc 3.2 (850)
5219 - failed open()s could cause deadlock; fixed (867, 869)
5220 - stop cancelling OST locks when files are closed (481)
5221 - overlapping XID spaces caused network corruption (851, 853)
5222 - fix unsafe fsfilt counter arithmetic; change to atomic_t
5223 - setattr_raw added, to do single-RPC, server-side setattrs
5224 - lmc/lconf syntax change for OST UUIDs
5225 - fix crashy race condition between ptlrpc_free_req and osc_close
5226 - don't use request in mdc_enqueue if we hit a timeout (889)
5227 - don't set the inode i_size for regular files from the MDS (896)
5228 - handle out of order completion AST (842)
5229 - don't LBUG if a lock request times out after receiving AST (913)
5230 - avoid d_rehash race in ll_find_alias by rehashing inside dcache_lock
5231 - if a bad lock AST arrives, send an error instead of dropping entirely
5232 - return 0 from revalidate2 if ll_intent_lock returns -EINTR (912)
5233 - fix leak in bulk IO when only partially completed (899, 900, 926)
5234 - fix O_DIRECT for ia64 (55)
5235 - (almost) eliminate Lustre-kernel-thread effects on load average (722)
5236 - C-z after timeout could hang a process forever; fixed (977)
5238 - client-side I/O cache (678, 924, 929, 941, 970)
5240 - READPAGE and SETATTRs which don't take server-side locks get
5243 2003-02-11 Phil Schwan <phil@clusterfs.com>
5246 - Fix ldlm_lock_match on the MDS to avoid matching remote locks (592)
5247 - Fix fsfilt_extN_readpage() to read a full page of directory
5248 entries, or fake the remainder if PAGE_SIZE != blocksize (500)
5249 - Avoid extra mdc_getattr() in ll_intent_lock when possible (534, 604)
5250 - Fix imbalanced LOV object allocation and out-of-bound access (469)
5251 - Most intent operations were removed, in favour of a new RPC mode
5252 that does a single RPC to the server and bypasses most of the VFS
5253 - All LDLM resource ID arrays were removed in favour of ldlm_res_id
5254 - Aggressively cancel local locks on DLM servers
5255 - mds_reint_unlink sends EA to the client if it's the last nlink.
5256 client uses that EA to unlink OST objects.
5257 - mds_reint_{rename,unlink,link} were rewritten to take ordered locks
5258 - recursive symlinks were fixed (439)
5259 - fixed NULL deref in DEBUG_REQ
5260 - filter_update_lastobjid no longer calls sync, which annoyed extN
5261 - fixed multi-client small-writes to a single file problem (445)
5262 - fixed mtime updates during file writes (607)
5263 - fixed vector writes on obdfilter causing problems when ENOSPC (670)
5264 - fixed bug in obd_brw_read/write() (under guise of testing 367)
5265 - fixed Linux OST size reporting problem (444, 656)
5266 - OST now updates object mtime with writes or setattr (607, 619)
5267 - client verifies file size before zeroing page past EOF (445)
5268 - OST now writes last allocated objid to disk with allocation (108)
5269 - LOV on echo now works (409)
5271 - mds_reint_unlink sends a new buffer, with the EA included. this
5272 buffer is only valid if body->valid & OBD_MD_FLEASIZE, which is only
5273 set if a regular file was being unlinked, and it was the last link
5274 - use PtlGet from the target for bulk writes (315)
5275 - OST now updates object mtime with writes or setattr (607, 619)
5276 - LDLM now has a grant-time callback to revalidate locked items, if
5278 - Many MDS operations were reorganized to combat race conditions
5280 - Merge b_intel branch (updated lprocfs code) - now at /proc/fs/lustre
5281 - configure check to avoid gcc version 2.96 20000731-2.96-98) (606)
5283 2003-01-06 Andreas Dilger <adilger@clusterfs.com>
5286 - Fully reactivate OST imports after reconnection (512, others)
5287 - Make sure client sees our -ENOTCONN from mds_handle (513 - partial)
5288 - More graceful error handling for truncating on dead OST (515)
5289 - Don't error out unless we're actually accessing dead stripes (474)
5290 - Fix garbage sizes when stripes are missing (410)
5291 - LRU counters were broken, causing constant lock purge (433, 432)
5292 - garbage on read from stripes with failed OSTs (441)
5293 - mark OSCs as active before reconnecting during recovery (438)
5294 - lov_enqueue and lov_cancel need to handle inactive OSTs (403)
5295 - lfind did not preserve OST order in output (443)
5296 - symlinks cause hung clients, incorrect data (439)
5297 - stop dereferencing request after dropping refcount (457)
5298 - don't LASSERT(spin_is_locked) on non-SMP (455)
5299 - fixes for many rename() bugs
5300 - fstat didn't correctly synchronize attributes (399)
5301 - server must handle lock cancellation during blocking AST prep (487)
5302 - bulk descriptors were free()d too soon (511)
5303 - fix paths in lconf, which would load incorrect modules (451, 507)
5304 - fix confusing lconf 'host not found' error message (386)
5305 - fix lock order deadlock on OST (O/R i_sem before journal ops, 478)
5306 - fix race condition in mdc_blocking_ast() for inode access (526)
5307 - fix lov_unpackmd() unpacking wrong number of stripes (537)
5308 - fix lov_set_osc_active() marking wrong OSC inactive (440)
5309 - fix bad lstripe lov_unpackmd() assertion (fix layering too) (527)
5310 - fix multiple writes of stripe MD to MDS (358, maybe 519)
5311 - fix lstripe in several ways (kernel side) (527)
5312 - fix request leak in ldlm_cli_enqueue (262)
5313 - incorrect OSC was marked inactive after OST failure
5314 - call mds_fs_cleanup before unmounting filesystem (524)
5315 - fix races between taking ns_lock and ldlm_lock_change_resource
5316 - fix races updating LOV export open file list
5317 - fix lov_enqueue error path, avoid decref-ing bad lock handle
5318 - fix recovery NULL deref in ldlm_cli_cancel_unused
5319 - fix some DLM races by using new hash table for lock handles (419)
5320 - permit the client to specify desired inodes, at replay
5321 - duplicate requests when we queue them for replay reintegration
5322 - fix last_rcvd offset calculation
5323 - sync after each recovered transaction, so we always make progress
5324 - never, not always, ERESTART requests without transnos
5325 - store the lov_desc in the MDS, so we don't depend on getlovinfo to
5327 - skip replay if the MDS says that the client is already connected
5328 - don't check for a recovery-enabled export to match lctl's UUID
5329 - don't INC_USE_COUNT for phantom exports
5330 - don't crash when cleaning up phantom exports (567)
5331 - don't double-finish or set replay data for errored mdc_open requests
5332 - abort requests when they time out, so we don't get old replies
5333 - send/receive replies for AST messages again
5334 - if the client says that it doesn't have the lock, cancel it on the
5336 - if we timeout during I/O, don't try to cancel an in-use lock; instead
5337 mark it as destroyed, it will all work out when decref is called
5338 - fix module use counts (22, 581)
5340 - ASTs now expect a reply (server cancels lock on error reply)
5342 2002-12-02 Andreas Dilger <adilger@clusterfs.com>
5345 - fix many simultaneous client startup (392)
5346 - fix dentry->d_it clobbering
5347 - credentials weren't being shipped for readdir/getattr operations
5348 - remove invalid assertions triggered during some concurrent MD
5350 - proper Lustre versions added (336, 389)
5351 - fix memory leak for create error case (398)
5352 - fix LOV locking bug that would get cli/srv out of sync
5353 - fix echo client over LOV (409)
5354 - fix dbench 2, extN refcount problem (170, 258, 356, 418)
5355 - fix double-O_EXCL intent crash (424)
5356 - avoid sending multiple lock CANCELs (352)
5358 - MDS can do multi-client recovery (modulo bugs in new code)
5360 - many updates, edits, cleanups
5362 2002-11-18 Phil Schwan <phil@clusterfs.com>
5365 - fix null d_it dereference (346)
5366 - fix full OST/dbench hang (333)
5367 - fix permission problem with file removal (286)
5368 - fix removal of OSCs from LOV when they fail
5369 - fix NULL deref during bulk timeout (214)
5370 - fix problems related to multiple filesystems on one MDS (241)
5371 - fixed serious subtle metadata locking bugs
5372 - free locks on clients when inodes are removed due to memory
5374 - fix inode pointer in lock data (285)
5375 - partial support for multiple MDS on a single host (241)
5376 - data locks weren't cancelled at clear_inode time (290, 311)
5377 - intent locks could lead to unbounded lock growth (205)
5378 - added a maximum lock count, an LRU list, and a flusher
5379 - fix multiple rename (365)
5380 - properly abstracted the echo client
5381 - OSC locked 1 byte too many; fixed
5382 - rewrote brw callback code:
5383 - fixed recovery bugs related to LOVs (306)
5384 - fixed too-many-pages-in-one-write crash (191)
5385 - fixed (again) crash in sync_io_timeout (214)
5386 - probably fixed callback-related race (385)
5388 - Add capability to MDS protocol
5389 - LDLM cancellations and callbacks on different portals
5391 2002-10-28 Andreas Dilger <adilger@clusterfs.com>
5394 - limit client IOV size to PTL_MD_MAX_IOV (611336, 191)
5395 - defer open object destruction to close time (601981, 138)
5396 - open/close OST file handle in obdo (OBD_MD_FLHANDLE) (601981, 138)
5397 - move LDLM_ENQUEUE/CONVERT back to MDS portal (625069)
5398 - abstract ll_lookup2, fix ll_revalidate2 to use abstraction (256)
5399 - don't call obd_setattr in ll_file_release for destroyed objects
5400 * protocol change to lustre_msg: move |version| and add |flags|
5401 * protocol change to osc_punch: "start" in "o_size", "end" in "o_blocks"
5402 * lock replay: for LDLM_FL_REPLAY trust client to do right thing
5403 * added replay of create, unlink, link and rename operations during
5404 MDS failover; recovery should be much more robust now
5405 * remove failed OSCs from LOVs (only lov_create uses this so far)
5406 * the lustre-HOWTO was brought (more) up to date (582544)
5408 2002-10-23 Phil Schwan <phil@clusterfs.com>
5411 - in-use dentries weren't being reused properly (617851)
5412 - prevent multiple LDLM setup (599178)
5413 - fix LOV size calculations for truncate (617853)
5414 - fix client handling of MDS intent errors (POSIX)
5415 - fix permission bug in lovstripe.c test (624321)
5416 - fix MDS thread deadlock - move LDLM handler to DLM portal (625069)
5417 - truncate past end of file could corrupt data
5418 - proper cleanup after timeouts, crashes, etc (592524, 550815)
5419 - a race in recovery could return ETIMEDOUT to apps (623947)
5420 - building outside the source directory was fixed
5421 * the lustre-HOWTO was brought (more) up to date (582544)
5422 * major progress was made on recovery functionality
5424 2002-10-10 Phil Schwan <phil@clusterfs.com>
5427 - recovery deadlock fix
5428 - rm -rf causes LBUG fix (617817)
5429 - file open by multiple tasks fix (618962)
5430 - directory permissions bugs (602707 and 620007)
5431 - journal_stop fixed with locking (611313)
5432 - O_APPEND failures resolved (618273, perhaps 614459)
5433 - lconf PATH fix (619770)
5434 - IA64 build fix (621450)
5435 - RPC buffer sizes scale with amount of memory
5437 2002-10-01 Phil Schwan <phil@clusterfs.com>
5440 - locks would be cancelled without throwing away data pages,
5441 resulting in inconsistent data (605627)
5442 - inode attributes were not always being refreshed (605627, 612449)
5443 - lconf now continues to cleanup after lctl reports an error
5444 - MDS now enforces user permissions (602707)
5445 - lprocfs cleanup fixed, but not yet enabled (614157)
5446 - fixed infinite server hang, should a client not respond to an AST
5447 - avoid going into recovery if user calls readlink() with a buffer
5448 that's too small (613941)
5449 - AST RPCs no longer require replies (614867) -- this may be changed
5450 - don't crash server if client sends an IOV that's too big (611336)
5451 - fixed lock conversion deadlock (611892)
5452 - fixed the following of symlinks (614622)
5453 * recovery: the server can remove locks from a client that dies, other
5454 clients can make progress
5455 * more extN patch fixes
5456 * compile-time configurable ptlrpc buffer allocations
5458 - collaborative read cache document
5459 - Lustre Lite Performance CDR document-in-progress
5461 2002-09-20 Andreas Dilger <adilger@clusterfs.com>
5464 - fix typo in patch-2.4.18
5466 2002-09-20 Andreas Dilger <adilger@clusterfs.com>
5469 - clear ptlrpc request each time in handle_incoming_request()
5470 - unlink of files now destroys the object on the OST
5472 2002-09-19 Peter Braam <braam@clusterfs.com>
5474 * add hard link support
5475 * change obdfile creation method
5476 * kernel patch changed
5478 2002-09-19 Peter Braam <braam@clusterfs.com>
5481 - stack overflow bug in extN fixed
5483 2002-09-18 Andreas Dilger <adilger@clusterfs.com>
5485 * documentation updates
5486 - add man pages for config tools
5487 - update tests/README to describe testing with new config tools
5488 - finish metadata API descriptions
5489 * bug fixes and cleanups
5490 - statfs workaround for 16TB limit
5491 - LOV stripe allocation improved, can stripe on subset of OSTs
5492 - LOV file size/IO offset was wrong for files > 4GB in size
5493 - object EA data was being dropped, caused files to be unreadable
5494 - memory overflow with non-LOV OST caused memory corruption
5495 - fixed regression tests to work with new config tools, obdfilter
5496 - fixed bug when directory size became larger than 1 block
5497 - fixed bug (for single client case) when PWD was deleted
5498 - invalidate local directory pages when doing intent-based ops
5499 - avoid LDLM oops when lock callback contained bad data
5501 2002-09-09 Andreas Dilger <adilger@clusterfs.com>
5503 * documentation updates
5504 * bug fixes and cleanups
5505 - configuration tools
5508 - 64-bit compile warnings
5509 - 64-bit internal statfs data
5511 * test_brw on persistent OST devices
5513 * lprocfs (disabled)
5515 2002-09-04 Andreas Dilger <adilger@clusterfs.com>
5517 * documentation updates
5518 * bug fixes and cleanups
5519 * configuration tools
5521 2002-08-30 Peter J. Braam <braam@clusterfs.com>
5524 * many small fixes to 0_5_4
5525 * io/network handling
5526 * thinkos in MDS operations
5528 2002-08-24 Peter J. Braam <braam@clusterfs.com>
5531 * crucial basic fixes to 0.5.3
5532 * IOR, Iozone work over Elan
5535 2002-08-07 Phil Schwan <phil@clusterfs.com>
5536 * version 0_5_3, our first alpha
5537 * we use the new Portals iovs
5538 * documentation updates
5539 * bug fixes and cleanups
5540 * small changes in the DLM wire protocol
5542 2002-07-25 Peter J. Braam <braam@clusterfs.com>
5543 * version 0_5_1 with some initial stability,
5544 * locking on MD and file I/O.
5545 * documentation updates
5546 * several bug fixes since 0.5.0
5547 * small changes in wire protocol
5549 2002-07-18 Phil Schwan <phil@clusterfs.com>
5551 * delivered as Lustre Light Alpha
5552 * fixed a crash after handling invalid MDS requests
5553 * fixed directory pages for architectures with non-4k pages sizes
5555 2002-07-11 Andreas Dilger <adilger@clusterfs.com>
5556 * release version v0_4_4
5557 * Moves TCP acceptor to be on port 2432 (unused Coda port) instead
5559 * Fixes a number of interruption problems with OST operations.
5560 * Update documentation for portals header changes
5561 * Move all wire protocol structs/defines to lustre_idl.h
5562 * Fixes symlink length bug.
5563 * Add tcpdump to repository.
5565 2002-07-05 Andreas Dilger <adilger@clusterfs.com>
5566 * release version v0_4_3
5567 * Fixes statfs for inodes on extN.
5568 * Fixes bug in runtests which would delete /etc/hosts.
5569 * Use 64-bit object IDs wherever possible (not into VFS though)
5570 Remove ost_get_info, which is unused by lustre, and out of date.
5572 2002-07-03 Peter Braam <braam@clusterfs.com>
5573 * release version v0_4_2 Fixes a lookup error (type not passed)
5574 * move forward to head of Portals
5575 * move forward to latest Lustre kernel
5577 2002-06-25 Peter Braam <braam@clusterfs.com>
5578 * release version v0_4_1. Hopefully stable on single node use.