Whamcloud - gitweb
fs/lustre-release.git
18 years ago- ph_subsys can be 7 symbols in HEAD
alex [Fri, 12 Aug 2005 19:01:54 +0000 (19:01 +0000)]
- ph_subsys can be 7 symbols in HEAD

18 years agodo reverse idmapping for remote perm fetch.
ericm [Fri, 12 Aug 2005 18:52:50 +0000 (18:52 +0000)]
do reverse idmapping for remote perm fetch.

18 years agofix someone's typo.
ericm [Fri, 12 Aug 2005 17:36:56 +0000 (17:36 +0000)]
fix someone's typo.

18 years ago- two ASSERTs int filter_parent() changed to ASSERTF with failed values for debug...
yury [Fri, 12 Aug 2005 15:15:33 +0000 (15:15 +0000)]
- two ASSERTs int filter_parent() changed to ASSERTF with failed values for debug purposes.

18 years ago- expanded error message in lov md sanity check function.
yury [Fri, 12 Aug 2005 13:29:02 +0000 (13:29 +0000)]
- expanded error message in lov md sanity check function.

18 years ago- fixed memory leak in ll_setxattr_internal()
yury [Fri, 12 Aug 2005 12:36:43 +0000 (12:36 +0000)]
- fixed memory leak in ll_setxattr_internal()

18 years ago- many fixes about using ENTRY, RETURN, GOTO and EXIT.
yury [Fri, 12 Aug 2005 12:21:14 +0000 (12:21 +0000)]
- many fixes about using ENTRY, RETURN, GOTO and EXIT.
- md_close() chnaged to use mdc_op_data like other md_* functions instead of using obdo.
- fixed prototype declaration for llapi_lov_get_uuids() - bug #6900
- fixed error message in ll_gns_mount_object() - bug #7146

- a lot of fixes about memory allocation/freeing:
  - checks is memory is actually allocated by OBD_ALLOC() before sing it and do appropriate
    actions otherwise.

  - fixed many leaks on error paths.
  - fixed missed OBD_FREE(capa_hash) in capa_cache_cleanup()

- fixed few on stack allocatoins of mdc_op_data in new landed stuff.

- in smfs_mount_cache() fixes correct cleanup in the cases of memory allocation failures.
  Do mntput() for new mounted cache_mnt on error path.

- in smfs_umount_cache() nullify smb->smsi_cache_ftype and smb->smsi_ftype after freeing
  them, to prevent attempt to free again if smfs_umount_cache() is called second time.

18 years agorevert previous commit: only transfer one record once time
niu [Fri, 12 Aug 2005 08:08:15 +0000 (08:08 +0000)]
revert previous commit: only transfer one record once time

18 years agomore of b7300: fix (actually hack around) the getattr reply buffer
ericm [Fri, 12 Aug 2005 07:12:41 +0000 (07:12 +0000)]
more of b7300: fix (actually hack around) the getattr reply buffer
positioning: make remote acl always occupy 2 buffers, just like local
acl.

18 years ago- don't call obd_set_info() on the NULL mdc export on rename.
niu [Fri, 12 Aug 2005 06:17:18 +0000 (06:17 +0000)]
- don't call obd_set_info() on the NULL mdc export on rename.
- transfer only one audit record on one notify.
- make sure the connection has been established before using the mdc export on name reconstruction.

18 years agoBranch HEAD
wangdi [Fri, 12 Aug 2005 04:55:41 +0000 (04:55 +0000)]
Branch HEAD
add type check in set_crypt

18 years agoBranch: HEAD
wangdi [Fri, 12 Aug 2005 03:40:39 +0000 (03:40 +0000)]
Branch: HEAD
fix typo usage for set_crypt

18 years agoadd some error message, try to catch the occasional lsd upcall failure.
ericm [Thu, 11 Aug 2005 23:47:58 +0000 (23:47 +0000)]
add some error message, try to catch the occasional lsd upcall failure.

18 years agoBranch HEAD
adilger [Thu, 11 Aug 2005 20:03:10 +0000 (20:03 +0000)]
Branch HEAD
Disable hlist macros unless compiling liblustre.  In b1_4 there are compat
macros in lustre_compat25.h that handle difference between 2.4 and 2.6
kernels instead of defining replacement macros.

18 years agosuppress a unnecessary message.
ericm [Thu, 11 Aug 2005 18:09:48 +0000 (18:09 +0000)]
suppress a unnecessary message.

18 years agofix regression introduced by capabilities.
ericm [Thu, 11 Aug 2005 18:06:50 +0000 (18:06 +0000)]
fix regression introduced by capabilities.
b=7300

18 years ago- fixes for making liblustre at least buildable (including fixes to compilation bug...
yury [Thu, 11 Aug 2005 11:55:33 +0000 (11:55 +0000)]
- fixes for making liblustre at least buildable (including fixes to compilation bug found by Brian recently)

18 years ago- fixes needed for building liblustre
yury [Thu, 11 Aug 2005 11:52:47 +0000 (11:52 +0000)]
- fixes needed for building liblustre

18 years agoBranch: HEAD
wangdi [Thu, 11 Aug 2005 08:45:05 +0000 (08:45 +0000)]
Branch: HEAD
some fixes about crypto api
1)some fixes about init inode key
2)some fixes about compute new mac

18 years ago- set only LOV_EA in mds_reint_setattr() in the case of CMOBD reint as setting sid...
yury [Thu, 11 Aug 2005 07:32:52 +0000 (07:32 +0000)]
- set only LOV_EA in mds_reint_setattr() in the case of CMOBD reint as setting sid and pid will confuse master MDS.
- set mds_security to "null" in lustre_connect_mds() if not specified or panic is possible.

18 years agoLand b_release_1_4_4 onto HEAD (20050810_0253)
adilger [Wed, 10 Aug 2005 17:23:16 +0000 (17:23 +0000)]
Land b_release_1_4_4 onto HEAD (20050810_0253)
Check for error writing to acceptor $PIDFILE.
Accptor reports when it has exited to syslog, so we can trace if it is a
culprit in the "bind: address already in use" problem.
b=5509
b=7283

18 years agoBackout "liblustre now builds" commit. It breaks building in utils:
brian [Wed, 10 Aug 2005 17:16:58 +0000 (17:16 +0000)]
Backout "liblustre now builds" commit.  It breaks building in utils:
  ../../lustre/include/linux/lustre_sec.h:541: error: field `c_hash' has incomplete type

18 years agoBranch: HEAD
wangdi [Wed, 10 Aug 2005 14:08:05 +0000 (14:08 +0000)]
Branch: HEAD
some fixes about crypto api

18 years ago- many fixes in liblustre related stuff. By now liblustre is at least build-able...
yury [Wed, 10 Aug 2005 13:03:33 +0000 (13:03 +0000)]
- many fixes in liblustre related stuff. By now liblustre is at least build-able in HEAD.

18 years ago- check connection before calling obd_set_info to set ids on an specified mdc.
niu [Wed, 10 Aug 2005 10:17:59 +0000 (10:17 +0000)]
- check connection before calling obd_set_info to set ids on an specified mdc.
- typo fix in mds_set_info.

18 years agoBranch HEAD
adilger [Wed, 10 Aug 2005 07:05:06 +0000 (07:05 +0000)]
Branch HEAD
Bump keepalive interval to send fewer total keepalives.
Trying to determine if this is cause of BG/L EMAC "flood" issue at
cleanup time.

18 years ago- rc code is not failure is > 0
tappro [Tue, 9 Aug 2005 22:42:00 +0000 (22:42 +0000)]
- rc code is not failure is > 0
- only files on OSS should be audited
- audit_fill_id_rec is valid only on mds. Moved to mds_audit.c
- audit_ost_get_id() gets lustre id for OSS object for audit log

18 years ago* #ifdef for extra proc_dointvec parameter under linux 2.6.8 in ranal and
eeb [Tue, 9 Aug 2005 15:24:44 +0000 (15:24 +0000)]
*  #ifdef for extra proc_dointvec parameter under linux 2.6.8 in ranal and
   openibnal (bz 7217)

*  remove WIRE_ATTR from iibanl kib_md_t (it's not a wire struct)

18 years ago* fixed openibnal/iibnal wire structs to use WIRE_ATTR. This (for openibnal)
eeb [Tue, 9 Aug 2005 14:50:26 +0000 (14:50 +0000)]
*  fixed openibnal/iibnal wire structs to use WIRE_ATTR.  This (for openibnal)
   is part of a fix for bz 7217

18 years agoBranch HEAD
wangdi [Tue, 9 Aug 2005 14:26:52 +0000 (14:26 +0000)]
Branch HEAD
get the correct uid, gid and mode when compute the new acl

18 years agoLand b_hd_capa onto HEAD (20050809_1942)
lsy [Tue, 9 Aug 2005 12:04:07 +0000 (12:04 +0000)]
Land b_hd_capa onto HEAD (20050809_1942)

18 years ago- put fill fid in a punch request
alex [Tue, 9 Aug 2005 10:55:17 +0000 (10:55 +0000)]
- put fill fid in a punch request

18 years agoBranch: b_hd_crypto
wangdi [Tue, 9 Aug 2005 10:12:37 +0000 (10:12 +0000)]
Branch: b_hd_crypto
some fixes about crypto api
1)some debug info fix
2)reorganize the code for mds_set_gskey

18 years ago- add mountoptions argument to lmv.sh
tappro [Tue, 9 Aug 2005 05:32:49 +0000 (05:32 +0000)]
- add mountoptions argument to lmv.sh

18 years ago- let's learn where uninitialized objects come from
alex [Mon, 8 Aug 2005 23:15:01 +0000 (23:15 +0000)]
- let's learn where uninitialized objects come from

18 years agob=7277
tappro [Mon, 8 Aug 2005 23:03:16 +0000 (23:03 +0000)]
b=7277

dentry is used before initialization.

18 years agob=4268
alex [Mon, 8 Aug 2005 22:57:45 +0000 (22:57 +0000)]
b=4268

 - additional debug info

18 years ago- add second mountpoint in the config
alex [Mon, 8 Aug 2005 22:55:08 +0000 (22:55 +0000)]
- add second mountpoint in the config

18 years ago- landed b_hd_mdref (mostly WB cache fixes)
yury [Mon, 8 Aug 2005 18:40:52 +0000 (18:40 +0000)]
- landed b_hd_mdref (mostly WB cache fixes)

18 years agoBranch: HEAD
wangdi [Mon, 8 Aug 2005 15:07:14 +0000 (15:07 +0000)]
Branch: HEAD
some minor fixes about sanity-crypto test scripts

18 years ago- another issue from landing day
tappro [Mon, 8 Aug 2005 14:37:01 +0000 (14:37 +0000)]
- another issue from landing day

18 years ago- root fid is 2 not 1
tappro [Mon, 8 Aug 2005 13:52:44 +0000 (13:52 +0000)]
- root fid is 2 not 1
- remove unneeded debug

18 years ago* Incorporated latest GMNAL updates from HP.
eeb [Mon, 8 Aug 2005 12:50:12 +0000 (12:50 +0000)]
*   Incorporated latest GMNAL updates from HP.
    Compiled but not run yet.

18 years agoBranch: HEAD
wangdi [Mon, 8 Aug 2005 09:26:28 +0000 (09:26 +0000)]
Branch: HEAD
some minor fixes about mks crypto

18 years agosplit dir after setting id in ea (in lfs dirstripe case)
niu [Mon, 8 Aug 2005 09:07:31 +0000 (09:07 +0000)]
split dir after setting id in ea (in lfs dirstripe case)
-b 7265

18 years agoneedn't get rootid while reconstruct pathname
niu [Mon, 8 Aug 2005 08:16:53 +0000 (08:16 +0000)]
needn't get rootid while reconstruct pathname

18 years ago- mds_validate_size() vs. mds_reint_unlink() race is absolutely legal
alex [Sun, 7 Aug 2005 16:53:26 +0000 (16:53 +0000)]
- mds_validate_size() vs. mds_reint_unlink() race is absolutely legal

18 years agob=7267,7034
alex [Sun, 7 Aug 2005 16:28:09 +0000 (16:28 +0000)]
b=7267,7034

 - dedicated threads for serving OST_DESTROY and MDS_CLOSE

18 years ago- don't mix var declarations and code
alex [Sun, 7 Aug 2005 16:14:49 +0000 (16:14 +0000)]
- don't mix var declarations and code

18 years ago- throw needless debugging bits away
alex [Sun, 7 Aug 2005 16:14:12 +0000 (16:14 +0000)]
- throw needless debugging bits away

18 years ago- couple minor optimizations
alex [Sun, 7 Aug 2005 14:23:39 +0000 (14:23 +0000)]
- couple minor optimizations

18 years ago- don't pass pid in case of slave creation.
alex [Sun, 7 Aug 2005 14:22:58 +0000 (14:22 +0000)]
- don't pass pid in case of slave creation.
  checking for non-zero fid is a wrong practice

18 years agob=7264
alex [Sun, 7 Aug 2005 11:18:14 +0000 (11:18 +0000)]
b=7264

 - mballoc must not be used during orphan cleanup

18 years ago- small fixes after landing
tappro [Sun, 7 Aug 2005 08:33:06 +0000 (08:33 +0000)]
- small fixes after landing

18 years ago- show n_link to learn the race
alex [Sun, 7 Aug 2005 07:11:03 +0000 (07:11 +0000)]
- show n_link to learn the race

18 years ago- minor fixes after landing day
alex [Sat, 6 Aug 2005 14:54:40 +0000 (14:54 +0000)]
- minor fixes after landing day

18 years ago- show correct id in debug line
alex [Sat, 6 Aug 2005 14:27:49 +0000 (14:27 +0000)]
- show correct id in debug line

18 years ago- return back original OST/MDS sizes
alex [Sat, 6 Aug 2005 14:25:33 +0000 (14:25 +0000)]
- return back original OST/MDS sizes

18 years ago- test 19 should take one more internal ea into account (parent fid)
alex [Sat, 6 Aug 2005 12:30:49 +0000 (12:30 +0000)]
- test 19 should take one more internal ea into account (parent fid)

18 years ago- missing back slash caused flood of complains from portals_debug_msg()
alex [Sat, 6 Aug 2005 12:19:37 +0000 (12:19 +0000)]
- missing back slash caused flood of complains from portals_debug_msg()

18 years ago- one more merging mess
alex [Sat, 6 Aug 2005 12:18:13 +0000 (12:18 +0000)]
- one more merging mess

18 years ago- typo from merging
alex [Sat, 6 Aug 2005 12:15:46 +0000 (12:15 +0000)]
- typo from merging

18 years ago- minor cleanups in that patch, quilt pop was claiming about old version
alex [Sat, 6 Aug 2005 12:15:15 +0000 (12:15 +0000)]
- minor cleanups in that patch, quilt pop was claiming about old version

18 years agolanding b_hd_sec, b6818, b7244
ericm [Sat, 6 Aug 2005 01:59:00 +0000 (01:59 +0000)]
landing b_hd_sec, b6818, b7244

18 years ago- b_hd_audit landing
tappro [Sat, 6 Aug 2005 01:06:14 +0000 (01:06 +0000)]
- b_hd_audit landing

18 years ago- minor correction
alex [Fri, 5 Aug 2005 23:31:06 +0000 (23:31 +0000)]
- minor correction

18 years ago- return right path back
alex [Fri, 5 Aug 2005 23:14:57 +0000 (23:14 +0000)]
- return right path back

18 years ago- some versions of python don't eat single member set w/o comma
alex [Fri, 5 Aug 2005 23:06:35 +0000 (23:06 +0000)]
- some versions of python don't eat single member set w/o comma

18 years agob=7039
alex [Fri, 5 Aug 2005 23:04:04 +0000 (23:04 +0000)]
b=7039

 - lov_change_cbdata() must iterate over all namespaces regardless activeness

18 years ago- track time recovery took and dump a log if it took >obd_timeout * 2
alex [Fri, 5 Aug 2005 22:45:06 +0000 (22:45 +0000)]
- track time recovery took and dump a log if it took >obd_timeout * 2
  (to simplify debugging tests 11 and 17)

18 years ago- additional debug in ldlm_lock_decref()
alex [Fri, 5 Aug 2005 22:42:43 +0000 (22:42 +0000)]
- additional debug in ldlm_lock_decref()

18 years ago- during last landing old version of the patch was checked in. fixed
alex [Fri, 5 Aug 2005 22:22:41 +0000 (22:22 +0000)]
- during last landing old version of the patch was checked in. fixed

18 years ago- code duplicate we got during landing b_hd_crypto
alex [Fri, 5 Aug 2005 22:09:12 +0000 (22:09 +0000)]
- code duplicate we got during landing b_hd_crypto

18 years agoAdd missing file from distribution.
brian [Fri, 5 Aug 2005 20:42:37 +0000 (20:42 +0000)]
Add missing file from distribution.

18 years ago- repair lmv.sh - it was broken for months
alex [Fri, 5 Aug 2005 20:34:14 +0000 (20:34 +0000)]
- repair lmv.sh - it was broken for months

18 years agocomment for ll_readahead_state and assertion in osc_send_oap_request() from b1_4
nikita [Fri, 5 Aug 2005 18:12:10 +0000 (18:12 +0000)]
comment for ll_readahead_state and assertion in osc_send_oap_request() from b1_4

18 years agoBranch: HEAD
wangdi [Fri, 5 Aug 2005 14:48:14 +0000 (14:48 +0000)]
Branch: HEAD
typo fix after land

18 years agoBranch: HEAD
wangdi [Fri, 5 Aug 2005 14:45:34 +0000 (14:45 +0000)]
Branch: HEAD
land crypto api

18 years ago- warn about the race in filter_intent_policy(), just to make sure it
alex [Fri, 5 Aug 2005 09:12:57 +0000 (09:12 +0000)]
- warn about the race in filter_intent_policy(), just to make sure it
  relates to errors mds gets sometimes

18 years agob=7039
alex [Thu, 4 Aug 2005 21:49:13 +0000 (21:49 +0000)]
b=7039

 - don't skip locks on inactive imports

18 years agob=7039
alex [Thu, 4 Aug 2005 19:14:37 +0000 (19:14 +0000)]
b=7039

 - dump additional debug info

18 years ago- missing \n
alex [Thu, 4 Aug 2005 13:32:51 +0000 (13:32 +0000)]
- missing \n

18 years agob=7039
alex [Thu, 4 Aug 2005 13:04:54 +0000 (13:04 +0000)]
b=7039

 - check for llite inode magic

18 years ago- no more flooding
alex [Thu, 4 Aug 2005 12:06:55 +0000 (12:06 +0000)]
- no more flooding

18 years agob=7039
alex [Thu, 4 Aug 2005 11:40:59 +0000 (11:40 +0000)]
b=7039

 - more debugging info

18 years agob=7049
alex [Thu, 4 Aug 2005 11:33:40 +0000 (11:33 +0000)]
b=7049

 - remove needless debugging checks

18 years agogive a warnning if somebody don't pass lustre id to ost, it should be
niu [Thu, 4 Aug 2005 09:45:49 +0000 (09:45 +0000)]
give a warnning if somebody don't pass lustre id to ost, it should be
error assertion in future.

18 years agoFix of careless (miss ;;)
liangzhen [Thu, 4 Aug 2005 09:00:13 +0000 (09:00 +0000)]
Fix of careless (miss ;;)

18 years agostore lustre id in ost object's EA.
niu [Thu, 4 Aug 2005 08:46:29 +0000 (08:46 +0000)]
store lustre id in ost object's EA.

18 years agoCFLAGS for compiling in Tiger
liangzhen [Thu, 4 Aug 2005 07:56:21 +0000 (07:56 +0000)]
CFLAGS for compiling in Tiger

18 years agob=7049
alex [Wed, 3 Aug 2005 21:56:47 +0000 (21:56 +0000)]
b=7049

 - try to unmap buffers from bd-mapping

18 years agob=7049
alex [Wed, 3 Aug 2005 13:29:38 +0000 (13:29 +0000)]
b=7049

 - it's enough to check for block number

18 years agob=7049
alex [Wed, 3 Aug 2005 12:18:42 +0000 (12:18 +0000)]
b=7049

 - it's better to unmap buffers holding no truncate sem
   note: unmap_underlying_metadata() is commented out to see whether
         a race happens at all

18 years agob=7049
alex [Wed, 3 Aug 2005 12:04:15 +0000 (12:04 +0000)]
b=7049

 - check for aliases in bd-mapping for writes only

18 years agob=7049
alex [Wed, 3 Aug 2005 10:20:08 +0000 (10:20 +0000)]
b=7049

 - let's check whether dirty pages/buffers in bd-mapping

18 years agob=7237
alex [Wed, 3 Aug 2005 07:34:22 +0000 (07:34 +0000)]
b=7237

 - missed wake_up() has been returned
 - few CDEBUG() have changed their level to D_ERROR to simplify debugging

18 years agoMerge b1_4_bug3389 from b1_4 (20050729_0312)
adilger [Tue, 2 Aug 2005 23:29:58 +0000 (23:29 +0000)]
Merge b1_4_bug3389 from b1_4 (20050729_0312)
Description: Mounting a Lustre file system on a node running as an OST could
             lead to deadlocks
Details    : OSTs now allocate memory needed to write out data at
             startup, instead of when needed, to avoid having to
             allocate memory in possibly low memory situations.
             Specifically, if the file system is mounted on on OST,
             memory pressure could force it to try to write out data,
             which it needed to allocate memory to do.  Due to the low
             memory, it would be unable to do so and the node would
             become unresponsive.
b=6514, b=5137

Description: Addition of lconf --service command line option
Details    : lconf now accepts a '--service <arg>' option, which is
             shorthand for 'lconf --group <arg> --select <arg>=<hostname>'
b=7015

Description: Failover mode is now the default for OSTs.
Details    : By default, OSTs will now run in failover mode.  To return to
             the old behaviour, add '--failout' to the lmc line for OSTs.
b=6101

Description: Health checks are now provided for MDS and OSTs
Details    : Additional detailed health check information on MSD and OSTs
             is now provided through the procfs health_check value.
b=1693

Description: Disk fragmentation on the OSTs could eventually cause slowdowns
             after numerous create/delete cycles
Details    : The ext3 inode allocation policy would not allocate new inodes
             very well on the OSTs because there are no new directories
             being created.  Instead we look for groups with free space if
             the parent directories are nearly full.
b=4466

Description: Network or server problems during mount may cause partially
             mounted clients instead of returning an error.
Details    : The config llog parsing code may overwrite the error return
             code during mount error handling, returning success instead
             of an error.
b=6302

18 years agob=3984
alex [Tue, 2 Aug 2005 19:46:23 +0000 (19:46 +0000)]
b=3984

 - a reference on resource must be taked under spinlock

18 years agob=7039
alex [Tue, 2 Aug 2005 18:46:35 +0000 (18:46 +0000)]
b=7039

 - concurrent lock cancelling might find inode dead while it's not yet

18 years agob=7231
alex [Tue, 2 Aug 2005 17:34:20 +0000 (17:34 +0000)]
b=7231

 - asynchronous commit in obdfilter/2.6