Whamcloud - gitweb
Robert Read [Mon, 28 Jun 2010 23:12:35 +0000 (16:12 -0700)]
Add missing .gitgnore files from new pthread directories.
Robert Read [Mon, 28 Jun 2010 18:04:28 +0000 (11:04 -0700)]
Revert "b=21128 run sync ldlm_bl_to_thread_list() in separate thread to save stack space."
This reverts commit
2cb1701012c4f1f0f1abdc99eb8ce6d4bb9c7846.
Still causing corruption, see comments on bug 21128.
Robert Read [Mon, 28 Jun 2010 18:04:00 +0000 (11:04 -0700)]
Revert "b=22244 delegate lock cancel to blocking thread"
This reverts commit
f96aff96fae1fcb3d06fd40ae4837ba975df12e3.
Reverting because of dependency on bug 21128.
Robert Read [Mon, 28 Jun 2010 18:00:18 +0000 (11:00 -0700)]
Revert "b=22244 handle list of locks in ldlm_bl_thread_main properly"
This reverts commit
f2360643cbb8136f97ec750547d12c30f88e0d4d.
Reverting because of dependency on bug 21128.
Fan Yong [Fri, 25 Jun 2010 22:14:22 +0000 (15:14 -0700)]
b=20581 Handle directory entry hash collisions
i=rread
i=pravin
Jian Yu [Fri, 25 Jun 2010 09:00:28 +0000 (17:00 +0800)]
b=19151 use single-level test directory for large-scale test 3a
$tdir is a two-level directory which causes the mode setting issue
while running MPI programs. Let's use a single-level test directory
for large-scale test 3a.
i=grev
i=robert.read
Vitaly Fertman [Thu, 24 Jun 2010 20:31:36 +0000 (00:31 +0400)]
b=22244 handle list of locks in ldlm_bl_thread_main properly
the fix to the original patch includes:
1. ldlm_bl_thread_main() cancels only locks left in the list after
ldlm_cli_cancel_list_local().
2. ldlm_bl_thread_main() and ldlm_cli_cancel_unused_resource() cancel
locks marked as LDLM_FL_BL_AST in the same RPC as others in the list.
i=green
i=johann
Vitaly Fertman [Thu, 24 Jun 2010 20:31:35 +0000 (00:31 +0400)]
b=22244 delegate lock cancel to blocking thread
i=adilger
i=green
re-landing: instead of cancelling locks locally in the shrinking thread,
deletate it to a separate blocking thread.
Nathan Rutman [Thu, 24 Jun 2010 17:49:30 +0000 (10:49 -0700)]
b=23004 bad cfs_user_write call causes oops
i=rread
i=breitz
Robert Read [Wed, 23 Jun 2010 16:30:24 +0000 (09:30 -0700)]
b=22911 Fix incompatibilies between ext4 and ext3 based ldiskfs
i=girish
i=rread
a=adilger
a=yangsheng
pravin [Thu, 24 Jun 2010 16:14:45 +0000 (21:44 +0530)]
b=21128 run sync ldlm_bl_to_thread_list() in separate thread to save stack space.
i=oleg.drokin
i=rahul
i=vitaly
Robert Read [Thu, 24 Jun 2010 17:02:12 +0000 (10:02 -0700)]
Update copyright messages.
Fan Yong [Thu, 24 Jun 2010 06:02:33 +0000 (14:02 +0800)]
b=21636 diagnostic patch to detect request delay send/reply
diagnostic patch to detect request delay send/reply
i=robert.read
i=pravin.shelar
Fan Yong [Thu, 24 Jun 2010 06:01:59 +0000 (14:01 +0800)]
b=19427 add the missing swab field for "rr_flags" in "mdt_reint_rec"
add the missing swab field for "rr_flags" in "mdt_reint_rec".
i=andreas.dilger
i=robert.read
pravin [Wed, 23 Jun 2010 08:45:01 +0000 (14:15 +0530)]
b=20563
i=zam
i=andreas
i=rahul
fix values of LUSTRE_SEQ_MAX_WIDTH and LUSTRE_SEQ_META_WIDTH.
this patch fixes fid to inode mapping so that it will be independent of
LUSTRE_SEQ_MAX_WIDTH.
there is fix for seq flush also.
Jian Yu [Wed, 23 Jun 2010 02:01:46 +0000 (10:01 +0800)]
b=22890 initialize MDSDEV in conf-sanity.sh
Initialize the MDSDEV variable in conf-sanity.sh.
i=grev
i=vladimir.saveliev
Robert Read [Mon, 21 Jun 2010 16:21:26 +0000 (09:21 -0700)]
b=22456 Remove obsolete patches and series files from ldiskfs/kernel_patches.
Brian J. Murrell [Tue, 22 Jun 2010 19:09:32 +0000 (15:09 -0400)]
b=22790 strip the .patch off of the dpatch name
When annotating the .dpatch files with the patch name, seeing as we
strip the .patch off and replace it with dpatch, do the same in the
annotation.
Brian J. Murrell [Tue, 22 Jun 2010 19:09:31 +0000 (15:09 -0400)]
b=20355 formalize support for pthreads
BG/P needs proper pthreads library arguments for linking, so properly
support pthreads properly in configure, using AX_PTHREAD.
The need for the pthread subdirs is because you cannot adjust the CC
variable value used on a per target basis, only for an entire Makefile.
This is the same problem we had with using the MPICC for some targets.
So to make some targets compile with the CC that the AX_PTHREAD macro
discovers/sets, we need to use a new Makefile. File locations remain
the same however.
i=panda (b1_8 version)
i=whitebear (b1_8 version)
Vladimir Saveliev [Mon, 21 Jun 2010 20:32:37 +0000 (00:32 +0400)]
b=21592 include linux/fs.h only when sys/mount.h doesn't define BLKGETSIZE64
i=andreas.dilger
i=andrew.perepechko
i=sheng.yang
Rahul Deshmukh [Mon, 21 Jun 2010 12:40:10 +0000 (18:10 +0530)]
b=20226 rename osd to osd-ldiskfs
adding .gitignore file and rename osd to osd-ldiskfs in test-framework.sh
Manoj Joseph [Fri, 18 Jun 2010 21:44:27 +0000 (15:44 -0600)]
b=20837 incomplete test output for ost-pools
Append output of create_perf_one() to stderr
i=robert.read
i=grev
Robert Read [Thu, 17 Jun 2010 22:34:47 +0000 (15:34 -0700)]
b=22168 Ignore EINTR returned from write
Robert Read [Fri, 18 Jun 2010 15:40:13 +0000 (08:40 -0700)]
Clean out more cruft.
Remove empty lustre/kernel-tests
Remove very outofdate README.kernel-source
Robert Read [Fri, 18 Jun 2010 15:33:59 +0000 (08:33 -0700)]
Remove remaining remnants of the CVS era.
Rahul Deshmukh [Thu, 17 Jun 2010 08:37:12 +0000 (14:07 +0530)]
b=20226 rename osd to osd-ldiskfs
i=rread
Nathan Rutman [Thu, 17 Jun 2010 18:23:03 +0000 (11:23 -0700)]
b=23094 add layout lock connect flag
o=jc.lafoucriere@cea.fr
i=nathan
i=rread
Vitaly Fertman [Thu, 17 Jun 2010 11:01:04 +0000 (15:01 +0400)]
b=22768 mdt open error handling
i=green
i=rread
mdt_finish_open() may return an error, which must result in unlocking
the object and returning it to client
Elena Gryaznova [Thu, 17 Jun 2010 10:23:30 +0000 (14:23 +0400)]
b=22040 use df POSIX output format
i=Oleg.Drokin
Rahul Deshmukh [Thu, 17 Jun 2010 05:57:02 +0000 (11:27 +0530)]
b=20226 rename osd to osd-ldiskfs
i=adilger
i=rread
Rename the osd to osd-ldiskfs
Rahul Deshmukh [Wed, 16 Jun 2010 06:51:43 +0000 (12:21 +0530)]
b=17569 add check in ext3-based ldiskfs for >8TB filesystems
i=johann
This is combined patch which includes both patches on bug 17569 which needs to
be landed.
1. remove force option for rhel5/ext4
2. add 16TB limit mount option for rhel5/ext4
Robert Read [Wed, 16 Jun 2010 16:45:11 +0000 (09:45 -0700)]
Revert "b=22755 Don't consume grant twice on recoverable resend"
This reverts commit
ccddc6afde64c75e39bb04fdeafec00c46f7f8e3.
Andreas Dilger [Tue, 15 Jun 2010 22:58:21 +0000 (16:58 -0600)]
b=16208 Report deprecated options, so they are removed from /etc/fstab.
This hunk of the patch was missed when the patch was originally landed.
Don't print "acl" in ll_show_options(), since it isn't needed at all.
Add #ifdef so the option parsing/messages are removed in the future.
o=sheng.yang
i=johann
i=andreas.dilger
Robert Read [Tue, 15 Jun 2010 23:40:05 +0000 (16:40 -0700)]
Update .gitignore files.
Robert Read [Wed, 9 Jun 2010 16:25:44 +0000 (09:25 -0700)]
b=22895 Cleanup replay-vbr test 7.
Add more messages so we can see what cycle we're in and
what op is being done.
i=tappro
di wang [Mon, 18 Jan 2010 18:58:47 +0000 (10:58 -0800)]
b=21379 Even create objects in DELORPHAN process; Choose osts in the pool in create remedy process
1.precreate the objects even in DELORPHAN process, but instead of allocating these objects, otherwise
it might create the gap of object ID, which might cause some orphan unlink log being recorded wrongly.
2.Only choose OSTs in the pool if in create remedy process.
i=adilger.dilger
i=robert.read
i=johann
Andrew Perepechko [Tue, 15 Jun 2010 20:08:27 +0000 (00:08 +0400)]
b=21665 "lfs osts" incorrectly tries to traverse lustre mount
We don't want to traverse the directory tree,
so return a positive value from sem_init to terminate
the traversal before it starts.
a=Joseph Herring (LLNL)
i=Andreas Dilger
i=Andrew Perepechko
Alexander.Zarochentsev [Tue, 15 Jun 2010 19:18:38 +0000 (23:18 +0400)]
b=22755 Don't consume grant twice on recoverable resend
a=oleg.drokin
i=eric.mei
i=johann
Alexander.Zarochentsev [Tue, 15 Jun 2010 19:00:15 +0000 (23:00 +0400)]
b=22755 check for OBD_CONNECT_GRANT_SHRINK in osc_should_shrink_grant
a=johann
i=oleg.drokin
i=di.wang
Alexander.Zarochentsev [Tue, 15 Jun 2010 19:00:14 +0000 (23:00 +0400)]
b=22755 Don't leak grant if write failed with EDQUOT
a=oleg.drokin
i=di.wang
i=johann
Alexander.Zarochentsev [Tue, 15 Jun 2010 19:00:13 +0000 (23:00 +0400)]
b=22755 Don't leak grant if shrink grant comes in write RPC
a=oleg.drokin
i=di.wang
i=andreas.dilger
i=johann
Robert Read [Tue, 15 Jun 2010 17:02:09 +0000 (10:02 -0700)]
Revert "b=22842 do not use bash variable to pass test setup state between sub-test"
This reverts commit
81fdcaed0df6041183240bd91f615e15baf9f087.
This causes a failure in test 101b.
Robert Read [Tue, 15 Jun 2010 15:58:10 +0000 (08:58 -0700)]
Revert "b=20563 Increase LUSTRE_SEQ_MAX_WIDTH"
This reverts commit
09803193a151902acc39720946b831b90655c4a8.
This patch causes several failures replay-vbr test 7.
Nicolas Williams [Mon, 14 Jun 2010 23:47:26 +0000 (18:47 -0500)]
b=22966 UID and GID "high" fields missing in a few structs
Several structs (mdt_body, mdt_remote_perm, obdo) needed
uid_h/gid_h fields for handling of SIDs (via FUIDs).
i=andreas.dilger@oracle.com
i=yong.fan@sun.com
i=johann@sun.com
Brian J. Murrell [Mon, 14 Jun 2010 17:12:29 +0000 (13:12 -0400)]
b=19336 fix landing error
Revert an inadvertant change leaked in between patch creation time
and landing time.
Liang Zhen [Sat, 12 Jun 2010 18:48:13 +0000 (11:48 -0700)]
b=22078 Bind ost/mdt to specific nids
The goal of the patch attached to this bug is to make NUMIOA factor management
easier, by giving the ability to restrict the NIDs that a target (mdt or ost)
registers on the MGS.
i=nathan
i=sebastien
Robert Read [Mon, 14 Jun 2010 23:47:44 +0000 (16:47 -0700)]
Update copyrights on source files changed since 2010-02-15.
Robert Read [Mon, 14 Jun 2010 23:37:33 +0000 (16:37 -0700)]
b=22990 Script to update copyright on recently modified source files
Based on the original script by Andreas.
i=adilger
i=johann
Manoj Joseph [Fri, 11 Jun 2010 20:40:45 +0000 (14:40 -0600)]
b=23040 Install lustre_rsync.8
Install the man page for lustre_rsync
i=robert.read
Robert Read [Fri, 11 Jun 2010 18:15:53 +0000 (11:15 -0700)]
Prepare for Build 44.
Brian J. Murrell [Fri, 11 Jun 2010 04:04:33 +0000 (00:04 -0400)]
b=22790 discover and utilize multiple CPUs
Brian J. Murrell [Fri, 11 Jun 2010 04:04:32 +0000 (00:04 -0400)]
b=22790 generate debian/patches
If we are running from a git checkout and there are patches on the
branch, (not counting the always-present debian/changelog update
patch, which is always assumed to be at the top of the patch stack)
put them into debian/patches/ as dpatches and remove them from the
source so that what's put into the source tarball is pristine.
Brian J. Murrell [Fri, 11 Jun 2010 04:04:31 +0000 (00:04 -0400)]
b=22790 fail make debs if autogen.sh is needed but not present
If a patch is added to debian/patches/ which modifies an autoconf
file but autogen.sh is not present, fail out of the "make debs"
build command.
Also provide a more explicit message to maintainers about what is
happening in this bit of code.
Nathan Rutman [Thu, 10 Jun 2010 23:30:17 +0000 (16:30 -0700)]
b=15253 fix conf-sanity 30b for non-tcp networks
i=rread
(also sync test with b18 version)
Brian J. Murrell [Thu, 10 Jun 2010 19:50:21 +0000 (15:50 -0400)]
b=19336 post landing cleanups
Remove generic find_linux_devel_paths()
- now that both the rhel5 and sles method files have their own particular
version of this method, remove this hacky-trying-to-work-for-both versions
from lbuild.
Remove a block of what is now redundant code.
Remove the comments from the target files describing what happened with
this bug.
Align the sles10 and sles11 target files:
- include the rpmfix specifier in the sles10 file
- remove the EXTRA_VERSION_DELIMETER from the sles10 file
- change the TARGET_DELIMETER to FLAVOR_DELIMETER in the sles11 file
Some whitespace cleanups.
i=wangyb
i=yangsheng
Brian J. Murrell [Thu, 10 Jun 2010 19:50:20 +0000 (15:50 -0400)]
b=19336 Get rid of the EXTRA_VERSION_DELIMETER shenanigans
We used to carry around a bunch of baggage in order to specify what kind
of delimeter to put between the version and "extra version". The truth of
the matter is that this should always be "-".
This patch includes some support for a build system developer to force
an uncached rebuild of all products.
i=wangyb
i=yangsheng
Vladimir Saveliev [Thu, 10 Jun 2010 03:07:56 +0000 (07:07 +0400)]
b=21244 utility to decode filter fid xattr
program and sanity test
i=andreas.dilger
Vladimir Saveliev [Thu, 10 Jun 2010 02:38:40 +0000 (06:38 +0400)]
b=21244 lustre_idl.h compilable from userspace
move few #include lines to make lustre/lustre_idl.h compilable from userspace
i=andreas.dilger
Vladimir Saveliev [Thu, 10 Jun 2010 01:42:57 +0000 (05:42 +0400)]
b=22842 do not use bash variable to pass test setup state between sub-test
tests 42, 101, 102 are fixed
i=robert.read
Vladimir Saveliev [Thu, 10 Jun 2010 00:57:23 +0000 (04:57 +0400)]
b=22828 fix unnormalized Linux symbols
introduced by patches of the bug 21485
i=robert.read
Alexander.Zarochentsev [Thu, 10 Jun 2010 10:35:05 +0000 (14:35 +0400)]
b=20498 print ldlm locks if obd_exports_barrier hangs.
i=tappro
i=robert.read
Alexander.Zarochentsev [Thu, 10 Jun 2010 10:35:04 +0000 (14:35 +0400)]
b=22598 Make COS lock inode bits aware
a=oleg.drokin
i=tappro
i=robert.read
Robert Read [Thu, 10 Jun 2010 06:48:53 +0000 (23:48 -0700)]
b=22850 add changelog entry
Johann Lombardi [Wed, 19 May 2010 12:07:49 +0000 (14:07 +0200)]
b=22850 bump max number of phys/hw segment in RHEL5 kernel
This grants us that we can always send 1MB i/o to the disk,
regardless of contiguous memory or not.
We already have the same tuning for SLES10.
Merge of 1.8 commits:
876c6e4d,
b0af14cf,
5d9a6cb7
Dmitry Zogin [Thu, 10 Jun 2010 05:58:17 +0000 (01:58 -0400)]
b=21563 Correcting the patch
Nicolas Williams [Wed, 9 Jun 2010 19:58:32 +0000 (14:58 -0500)]
b=23023 wirecheck.c / newwiretest don't build
wirecheck.c needed an update following b19427. Several *_ogr
struct fields became *_oseq.
i=andreas.dilger@oracle.com
Andreas Dilger [Fri, 4 Jun 2010 17:54:07 +0000 (11:54 -0600)]
b=22481 Add "lfs_migrate" script from manual into lustre/scripts and RPMs
Add enhanced lfs_migrate script from manual into lustre/scripts directory
and in RPM packages. It does a "poor man's" migration of files from
their current OST layout to a new OST layout as chosen by the MDS.
It is currently not safe to use for files that are being written to.
It will result in the current file becoming an open-unlinked file, and
the migrated file will be used for any subsequent opens. Any writes to
either the old or migrated file will in clients seeing different data,
and writes to the old file will be lost when it is closed.
There is no direct control of object allocation within the script, it
depends on external control to manage space. Use "lfs find" to
select appropriate files to migrate from full OSTs. "lctl --deactivate"
is needed on the MDS to deactivate OSTs if trying to empty an OST
completely, otherwise QOS will tend to migrate files to empty OSTs.
i=brian.murrell
i=nathan.rutman
pravin [Thu, 10 Jun 2010 03:55:37 +0000 (09:25 +0530)]
b=20563 Increase LUSTRE_SEQ_MAX_WIDTH
Increase value of LUSTRE_SEQ_MAX_WIDTH, fix fid_flatten() so that it would not
depend on LUSTRE_SEQ_MAX_WIDTH.
i=rahul
i=andreas
pravin [Thu, 10 Jun 2010 03:50:50 +0000 (09:20 +0530)]
b=20563 add cfs_curproc_is_32bit() function to check binary compatibility.
i=rahul
i=andreas
LiuYing [Thu, 10 Jun 2010 02:55:07 +0000 (10:55 +0800)]
b=22455 add list/get/set_param to lctl man page
-add list/get/set_param to lctl man page
-add list_param "-R" option to command usage
i=adilger
i=rread
Eric Mei [Thu, 10 Jun 2010 02:15:37 +0000 (20:15 -0600)]
b=22458 consolidate fs_db flags.
i=rread
i=adilger
Dmitry Zogin [Wed, 9 Jun 2010 22:00:21 +0000 (18:00 -0400)]
b=22747 Processes stuck waiting for inode I_LOCK bit to clear.
Use CFS_ALLOC_IO instead of CFS_ALLOC_STD in OBDO_ALLOC
i=andreas.dilger
i=johann
Dmitry Zogin [Wed, 9 Jun 2010 20:04:10 +0000 (16:04 -0400)]
b=21563 Metadata performance has degraded for some operations between 1.6.5 and 1.8.1
Remove the check for OSCC_FLAG_CREATING flag in osc_precreate(), so that
it enters oscc_internal_create(). Also remove the check for in
oscc_internal_create() for OSCC_FLAG_DEGRADED - let the chance to create
objects in case all OSTs are degraded.
i=andreas.dilger
i=johann
Rahul Deshmukh [Mon, 7 Jun 2010 06:31:55 +0000 (12:01 +0530)]
b=21951 2.6.32-fc13 patchless client support for HEAD
i=zhen.liang
Sebastien's patch to use sg_init_table for fc13 patchless client (2.6.32 kernel)
Robert Read [Thu, 10 Jun 2010 04:17:26 +0000 (21:17 -0700)]
Revert "b=19325 adjust waiting extent locks during 1st enqueue"
This reverts commit
c022258ea05eb2d5fecfd44d8b5ece3f1d8ec095.
Reverting this again because of data corruption issues seen in
bug 20824 (sanitn test_16).
Robert Read [Wed, 9 Jun 2010 16:42:34 +0000 (09:42 -0700)]
b=22960 Update mdsrate scripts to parse new format
i=nico
Robert Read [Tue, 8 Jun 2010 16:18:42 +0000 (09:18 -0700)]
Revert "b=21128 run sync ldlm_bl_to_thread_list() in separate thread to save stack space."
This reverts commit
f79f39bd69d40faea2bd06bd1c342430553e57b8.
Appears be related to bugs 22823 and 22715.
Eric Mei [Fri, 4 Jun 2010 23:21:14 +0000 (17:21 -0600)]
b=22520 make sure cl_inode_fini() to destroy the cl_object.
r=rread
r=wangdi
Nathan Rutman [Fri, 4 Jun 2010 22:53:53 +0000 (15:53 -0700)]
b=15253 add failover nidlist to proc import
i=adilger
i=rread
Andrew Perepechko [Fri, 4 Jun 2010 20:39:32 +0000 (00:39 +0400)]
b=22194 lfs quota output refinement
Added more dashes to lfs quota output to make it parseable with scripts.
Added -q option to lfs quota.
sanity-quota "lfs quota" calls were compactified.
a=Joseph Herring
a=Christopher Morrone
a=Andrew Perepechko
i=Johann Lombardi
i=ZhiYong Tian
i=Andrew Perepechko
Dmitry Zogin [Fri, 4 Jun 2010 18:31:39 +0000 (14:31 -0400)]
b=22889 OSS Kernel panic - Kernel BUG at fs/jbd/transaction.c:1155
extend truncate transaction to have enough blocks for last_rcvd and VBR
o=andreas.dilger
i=tappro
i=oleg.drokin
Robert Read [Fri, 4 Jun 2010 21:17:06 +0000 (14:17 -0700)]
b=22960 Add effective and avg client rates to mdsrate output.
i=adilger
Wang Di [Fri, 4 Jun 2010 15:36:30 +0000 (11:36 -0400)]
b=19427 correct lmm_object_id and reserve fids for fid-on-OST.
1. Change lmm_object_id to fid.
2. Cleanup fid spaces reservation (for fid-on-OST).http://arch.lustre.org/index.php?title=Interoperability_fids_zfs#NEW.0
3. Rename group to Seq.
i=Adilger
i=Pravin.shelar
i=Robert
Oleg Drokin [Fri, 4 Jun 2010 00:08:20 +0000 (20:08 -0400)]
b=20969 Doxygen comments for ptlrpc
Adds explanations to a lot of ptlrpc functions and structures
i=liang
Nathan Rutman [Wed, 2 Jun 2010 18:49:46 +0000 (11:49 -0700)]
b=20972 Add doxygen comments to MGS
i=breitz
Vladimir Saveliev [Tue, 1 Jun 2010 21:44:56 +0000 (01:44 +0400)]
b=13698 lfsck.sh support for remote mds and oss
SHARED_DIRECTORY is to be specified in case of remote servers
LFSCK_ALWAYS flag to check the filesystem from check_and_cleanup_lustre()
it is "no" by default
i=adliger
i=grev
LiuYing [Wed, 2 Jun 2010 10:29:46 +0000 (18:29 +0800)]
b=22754 a small bug in LDLM_POOL_PROC_WRITER
lprocfs_wr_unit returns "count >= 0" for success
i=nathan
i=fanyong
LiuYing [Wed, 2 Jun 2010 10:24:34 +0000 (18:24 +0800)]
b=22455 fix the problem in get_param with option "-N"
-set flag recursive=0 when we get param with option "-N"
-use operator "=" to set param with negative value
i=adilger
i=rread
hongchao.zhang [Thu, 27 May 2010 22:18:16 +0000 (06:18 +0800)]
b=21938 wait and signal correct waitq
update patch to use req->rq_set_waitq instead of req->rq_reply_waitq
i=johann
i=robert.read
Vitaly Fertman [Wed, 2 Jun 2010 09:19:33 +0000 (13:19 +0400)]
b=22864 'som_preview' mount option for lustre client.
i=adilger
i=tappro
a mount option 'som_preview' is added on the client side to send
SOM_CONNECT flag only if specified.
Rahul Deshmukh [Wed, 2 Jun 2010 06:04:54 +0000 (11:34 +0530)]
b=21399 ./acceptance-small.sh: line 109: 5638 Segmentation fault
i=tappro
Removed the constant of req_format and req_msg_field as they are causing seg
fault (on slell clients) while initializing in req_layout_init() .
Fan Yong [Wed, 2 Jun 2010 01:44:57 +0000 (09:44 +0800)]
b=18741 replace checking the return value of "lfs setquota" with checking the output message of it for detailed error
replace checking the return value of "lfs setquota" with checking the output message of it for detailed error
i=andrew.perepechko
i=robert.read
Fan Yong [Wed, 2 Jun 2010 01:30:17 +0000 (09:30 +0800)]
b=22560 fix OBD_CONNECT_FULL20 for conflict with OBD_CONNECT_MAX_EASIZE
fix OBD_CONNECT_FULL20 for conflict with OBD_CONNECT_MAX_EASIZE
i=robert.read
i=andreas.dilger
Fan Yong [Wed, 2 Jun 2010 01:16:05 +0000 (09:16 +0800)]
b=22853 test should stop at once if any dbench failed for replay-single test_70b
test should stop at once if any dbench failed for replay-single test_70b
i=jian.yu
i=robert.read
Jian Yu [Mon, 31 May 2010 01:30:02 +0000 (09:30 +0800)]
b=21097 work around for bash mangling of md5sum result
Store the md5sum results to a file and then have md5sum check the results
in the file.
o=James Simmons
i=jian.yu
i=edg
Robert Read [Fri, 28 May 2010 17:47:09 +0000 (10:47 -0700)]
Prepare for Build 43.
Manoj Joseph [Fri, 28 May 2010 00:42:32 +0000 (18:42 -0600)]
b=22298 Buffalo should be able to detect test failures
Buffalo should be able to detect test failures even when the test
does not use run_test.
i=robert.read
i=nicolas.williams
Minh Diep [Tue, 25 May 2010 20:31:51 +0000 (22:31 +0200)]
b=22509 increase default md stripe_cache_size to 16k
i=adilger
i=brian
Mikhail Pershin [Thu, 27 May 2010 19:55:29 +0000 (23:55 +0400)]
b=20997 Enable time limit on the request to be queued
i=rread,nathan
This feature is enabled by setting the time limit on the request. The limit
is on how long the request is queued, whether or not it's ever been sent.
IOW, it's not a "resend" limit, because it can expire even if the request
has never been sent the first time.
Patch removes also unused imp flags and modify conf-sanity test to use 'mds
mount first' scheme to avoid long test time
Alexander.Zarochentsev [Thu, 27 May 2010 11:02:28 +0000 (15:02 +0400)]
b=21140 Restore high priority request handling
i=vitaly.fertman
i=robert.read