Whamcloud - gitweb
Bobi Jam [Wed, 4 May 2011 11:35:30 +0000 (19:35 +0800)]
LU-273 Remove export-show_task patch
It is not used, dump_trace() is enough for the purpose.
Change-Id: Ic67ce634a8985835b00373024c88608fe0ef114f
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/493
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Tested-by: Hudson
hongchao.zhang [Sun, 3 Apr 2011 07:28:09 +0000 (15:28 +0800)]
LU-213 using direct io for quota test to avoid cache problem
in test_14a of sanity-qutota.sh, the data written was cached in client side,
which causes the test failed, using "mulitop ${TESTFILE} oO_DIRECT" (direct IO)
instead of "dd" to avoid such issue.
Change-Id: I0802c2ea200483c148e4e7f61e2c6d4b34500f01
Signed-off-by: Hongchao Zhang <hongchao.zhang@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/435
Tested-by: Hudson
Reviewed-by: Yu Jian <yujian@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Brian J. Murrell [Mon, 9 May 2011 19:11:12 +0000 (15:11 -0400)]
LU-278 only use the v* tag and limit the match
In order to allow people to create tags such as 2.0.59-1foobar,
only try to match on the v* type tags, and even then, tighten up
the match pattern (as best we can using globs) to try to limit the
number of false positives.
Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: I2d6e422fc0c17df7eb4169efea8e3557e80b941d
Reviewed-on: http://review.whamcloud.com/517
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Yang Sheng <ys@whamcloud.com>
Reviewed-by: Michael MacDonald <mjmac@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Christopher J. Morrone [Tue, 10 May 2011 00:49:45 +0000 (17:49 -0700)]
LU-298 Improve check for dpkg system.
Use lsb_release id (Ubuntu, Debian, RedHatEnterpriseServer, etc.)
to determine if the system uses dpkg, instead of just checking for
the existence of the dpkg binary. For example, some people do
install dpkg on their RHEL systems for various reasons.
Change-Id: I74db5277cb2db00927b74b652e7945bd95bca093
Signed-off-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-on: http://review.whamcloud.com/519
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Brian J. Murrell <brian@whamcloud.com>
Reviewed-by: Chris Gearing <chris@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Brian J. Murrell [Tue, 3 May 2011 19:57:21 +0000 (15:57 -0400)]
LU-303 fix ofed download for bugfix releases
On the OFED download server both w.x.y and w.x.y.z releases are
stored in the w.x.y dir, so handle that properly for the w.x.y.z case.
Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: Ic0ba6861ce000a9cc47c63219e357d2f6636b951
Reviewed-on: http://review.whamcloud.com/527
Tested-by: Hudson
Reviewed-by: Yang Sheng <ys@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Michael MacDonald <mjmac@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Brian J. Murrell [Tue, 3 May 2011 19:57:21 +0000 (15:57 -0400)]
LU-301 bump kernel version for sles11
We need to provide the same "self service" love in lbuild for sles11
(client only) building.
May as well bump the supported kernel to the latest while I'm at it.
Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: I84617f429b0cef23a24fe161158580917ffe3041
Reviewed-on: http://review.whamcloud.com/482
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Michael MacDonald <mjmac@whamcloud.com>
Reviewed-by: Chris Gearing <chris@whamcloud.com>
Reviewed-by: Yang Sheng <ys@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Prakash Surya [Tue, 10 May 2011 16:58:44 +0000 (09:58 -0700)]
LU-160 Split small/big file parts of test_155_load
Separate the small file parts and the big file parts of test_155_load
into different tests. This has the benefit of always running the small
file portions of test 155, even if the big file tests are skipped.
Signed-off-by: Prakash Surya <surya1@llnl.gov>
Change-Id: Ic895fb8590f1d15fbd1889b64b0a6143ff1c6289
Reviewed-on: http://review.whamcloud.com/528
Tested-by: Hudson
Reviewed-by: Yu Jian <yujian@whamcloud.com>
Reviewed-by: Chris Gearing <chris@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Jinshan Xiong [Wed, 11 May 2011 02:20:12 +0000 (19:20 -0700)]
LU-300: Oops in cl_page_put of page fault path
This problem is due to the unintialization of cl_io, so that an invalid page
address will be accessed in vvp_io_fault_fini().
Change-Id: Ic85a5d22e175cc74ada0e51831142d9f37798c08
Signed-off-by: Jinshan Xiong <jay@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/530
Tested-by: Hudson
Reviewed-by: wangdi <di.wang@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andreas Dilger [Fri, 29 Apr 2011 21:59:52 +0000 (15:59 -0600)]
LU-255: enable ext4 features by default
Enable the flex_bg, huge_file, and dir_nlink features from ext4 by
default for ext4-based ldiskfs filesystems. The flex_bg feature
can significantly reduce e2fsck time as well as time-to-first-write
on newly mounted OSTs.
Reduce the number of inodes created on larger OST filesystems, since
there are far too many inodes created by default, which wastes space
and slows down e2fsck.
Increase the number of inodes on MDT filesystems, subject to
constraints from the default number of stripes being stored for
each inode, which may increase the inode size being used. Remove
changes to the statfs() results on the MDT to limit inode count to
the returned block count since increased MDT inodes will always
exceed free blocks, and it was confusing.
Skip zeroing the journal at format time. It will be overwritten
at first use, and avoids writing 400MB of zeroes to the filesystem
needlessly. We can't skip zering the inode table by default, but
newer mke2fs will check if the kernel handles this itself, and
request it internally.
Cleanups to remove support for 2.4 kernels and reiserfs filesystem.
Change-Id: I519bffaad5e97ee68c189efb00e07ebc0b1600f5
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/480
Reviewed-by: Yu Jian <yujian@whamcloud.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
hongchao.zhang [Sat, 30 Apr 2011 01:27:20 +0000 (09:27 +0800)]
LU-276 increase file count to mitigate affect of random condition
in test_18 of ost_pools, increase the count of created file to
mitigate the affect of various random condition which could
lead to false result
Change-Id: Ieca04b37ccc66698cd2159fa1a83f2f9fe07dc41
Signed-off-by: HongChao Zhang <hongchao.zhang@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/550
Tested-by: Hudson
Reviewed-by: Mikhail Pershin <tappro@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Christopher J. Morrone [Tue, 10 May 2011 02:34:51 +0000 (19:34 -0700)]
LU-299 Initialize struct shrinker variable in test
If the "tmp" struct in the test is uninitialized, the test
will fail for that reason, rather its intended functions.
Change-Id: I4b1e97e511c748155f742909d9424f228fc24545
Signed-off-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-on: http://review.whamcloud.com/522
Tested-by: Hudson
Reviewed-by: Bobi Jam <bobijam@whamcloud.com>
Tested-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Bobi Jam [Tue, 10 May 2011 03:14:15 +0000 (11:14 +0800)]
LU-279 Test failure on test suite sanity, subtest test_17k
rsync 3.0.x cares about exact xattr value size when passing 0 size
for determining the buffer size to hold the xattr value.
Change-Id: I05f8c062e64e22dd6982dc752d5869b457c7a2ef
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/524
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-by: Prakash Surya <surya1@llnl.gov>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Bobi Jam [Tue, 3 May 2011 07:50:15 +0000 (15:50 +0800)]
LU-260 Add open-unlinked dir regression test case.
Change-Id: Ic9efd3d89377bc34470cb3b7ab88f62b2a03b058
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/485
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mikhail Pershin [Fri, 22 Apr 2011 18:24:25 +0000 (22:24 +0400)]
LU-128 Avoid assertion on wire data in last_rcvd update
- checks that lower transno can't overwrite the bigger one in last_rcvd
slot.
- evict client if bad transno was sent in replay, this is done by simulating
VBR failure
- keep assertion for server-generated transno, this is logical error
- fix issue with resent-replay open which can cause unexpected
transaction while closing the old mfd.
Change-Id: Ib523c25408b4d821f52d95c40a4fbd7d79d6cbe6
Signed-off-by: Mikhail Pershin <tappro@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/452
Tested-by: Hudson
Reviewed-by: Niu Yawei <niu@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Yu Jian [Mon, 28 Mar 2011 09:52:27 +0000 (17:52 +0800)]
LU-160 check the minimum available size of OST in test_155_load()
Fix test_155_load() in sanity.sh to check whether the minimum available
size of OST is bigger than the large file size. If not, then sanity test
155{a,b,c,d} would be skipped.
Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: I0a587337ad188a96f2ada51726604029c666df07
Reviewed-on: http://review.whamcloud.com/368
Tested-by: Hudson
Reviewed-by: Prakash Surya <surya1@llnl.gov>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
nasf [Wed, 4 May 2011 13:43:11 +0000 (21:43 +0800)]
LU-271 Skip permission check for NFS open after create
For open_create operation, NFS client will trigger separate open RPC after create RPC.
We only need to check create permission for such case.
Change-Id: I7f0e2edb1b14f1044790180345f2c8328d92d4e8
Signed-off-by: nasf <yong.fan@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/494
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Reviewed-by: Lai Siyao <laisiyao@whamcloud.com>
Oleg Drokin [Wed, 11 May 2011 03:06:44 +0000 (23:06 -0400)]
New version 2.0.60
Signed-off-by: Oleg Drokin <green@whamcloud.com>
Change-Id: Id062aed81b8956d42a8c5487f53bb44e505b491e
Ned A. Bass [Wed, 20 Apr 2011 21:56:09 +0000 (14:56 -0700)]
LU-225 --with-ldiskfs-inkernel checks wrong file
Commit
fd0bd149e66cbc06a17c372e159f9bfba024cc97 changed the file checked
when the --with-ldiskfs-inkernel configure option is used from
$LINUX/fs/ldiskfs/inode.c to $LINUX/include/linux/ldiskfs_fs.h.
It looks like the person fixing a similar case statement assumed the
same fix applied to the inkernel case. However, this is wrong because
the purpose of --with-ldiskfs-inkernel is to allow us to build lustre
without having full kernel sources available, so inode.c won't exist.
Signed-off-by: Ned Bass <bass6@llnl.gov>
Change-Id: I6dc1db41fdee0016eab7d151c519e3359d78b44b
Reviewed-on: http://review.whamcloud.com/446
Reviewed-by: Brian J. Murrell <brian@whamcloud.com>
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Yu Jian [Tue, 3 May 2011 12:55:30 +0000 (20:55 +0800)]
LU-235 fix sanity test 36g to work with remote OSS
Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: Iccf7f109c69ac7dd34f9f5b7ee2d5ce4cb892053
Reviewed-on: http://review.whamcloud.com/488
Tested-by: Hudson
Reviewed-by: Chris Gearing <chris@whamcloud.com>
Reviewed-by: Brian J. Murrell <brian@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
yangsheng [Fri, 6 May 2011 03:54:09 +0000 (11:54 +0800)]
LU-188 Reduce startup messages.
commit
f257c3021c7470445673bc930c4884c5df9b9429
Date: Mon Jan 24 15:17:35 2011 +0100
b=17275 make lustre client less verbose at startup time for Cray
i=johann
Signed-off-by: YangSheng <ys@whamcloud.com>
Change-Id: I78a5196f56ff147075676d2fdcdc6d9978437d72
Reviewed-on: http://review.whamcloud.com/406
Tested-by: Hudson
Reviewed-by: Fan Yong <yong.fan@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Brian J. Murrell [Thu, 5 May 2011 19:26:56 +0000 (15:26 -0400)]
LU-284 strip .$arch off of kern vers for requires
RHEL6 appends .arch to the end of the version in the utsrelease.h file.
The kernel RPM however provides that value with the .arch stripped off,
so do the same stripping in the lustre-modules RPM for it's Requires:.
Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: Ib6ff32c006f184edceb5d667b649bb106d503faf
Reviewed-on: http://review.whamcloud.com/502
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Tested-by: Hudson
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Bobi Jam [Tue, 26 Apr 2011 08:45:48 +0000 (16:45 +0800)]
LU-240 test_171() races with background multiop
Change-Id: Ic05c8d79e82cfc632942b13ec3b46ee26da49555
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/466
Reviewed-by: Li Wei <liwei@whamcloud.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mikhail Pershin [Thu, 21 Apr 2011 19:17:43 +0000 (23:17 +0400)]
LU-178 prevent replays after recovery on server
- Drop requests with wrong flags during normal processing.
- remove race in target_handle_connect which can cause replays over
new export
- test case for wrong replays handling
Change-Id: Icbfc0ae549ab63a510e0d62fcac1ed661b324fb9
Signed-off-by: Mikhail Pershin <tappro@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/391
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Reviewed-by: Niu Yawei <niu@whamcloud.com>
Lai Siyao [Wed, 27 Apr 2011 01:23:08 +0000 (18:23 -0700)]
LU-239 Lustre Header (include/lprocfs_status.h) is not ISO C99 complaint
Signed-off-by: Lai Siyao <laisiyao@whamcloud.com>
Change-Id: I40ce6ac0ecaaf19c01cd25603971ff1fb8ed10df
Reviewed-on: http://review.whamcloud.com/465
Tested-by: Hudson
Reviewed-by: Yang Sheng <ys@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
yangsheng [Wed, 27 Apr 2011 16:03:04 +0000 (00:03 +0800)]
LU-234 OOM killer causes node hang.
b=18213 Handle the signal to avoid process hang.
Change-Id: Ic730d98fb812bc9dbb5249847a6115a84d67758b
Signed-off-by: Yang Sheng <ys@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/470
Reviewed-by: Fan Yong <yong.fan@whamcloud.com>
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Bobi Jam [Tue, 3 May 2011 03:18:55 +0000 (11:18 +0800)]
LU-259 Wrong option in ext4-force_over_16tb-rhel6.patch
Change-Id: I0b35fc06b3674fd3f8b3334a2af8e8e5b346f910
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/484
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Prakash Surya [Thu, 5 May 2011 18:01:56 +0000 (11:01 -0700)]
LU-283 Clean up trailing whitespace in sanity.sh
The sanity.sh script has some trailing whitespace issues scattered about
the file. This patch is an attempt to clean these up.
Signed-off-by: Prakash Surya <surya1@llnl.gov>
Change-Id: Ie59fe322ec23ba9e2ce3006dc7d3e9d5ee3e24b5
Reviewed-on: http://review.whamcloud.com/501
Tested-by: Hudson
Reviewed-by: Chris Gearing <chris@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Lai Siyao [Thu, 5 May 2011 02:59:34 +0000 (19:59 -0700)]
LU-243 async lov_sync() operation
Port bz 17239 to 2.1: use RPC set for obd_sync().
Signed-off-by: Lai Siyao <laisiyao@whamcloud.com>
Change-Id: Ie139dd6a4c141f3309814655627112eed6e71b40
Reviewed-on: http://review.whamcloud.com/496
Tested-by: Hudson
Reviewed-by: Niu Yawei <niu@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Jinshan Xiong [Tue, 26 Apr 2011 05:31:52 +0000 (22:31 -0700)]
LU-77 cl_page.c::cl_page_own0() assertion in echoclient
There is no lock for echo_page, it is easy to hit this issue
when obdfilter_survey is doing IO in parallel.
Change-Id: I4bacb5a72afd7bad68bdabf5a64d711e34b5ea3c
Signed-off-by: Jinshan Xiong <jay@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/462
Tested-by: Hudson
Reviewed-by: Lai Siyao <laisiyao@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
yangsheng [Tue, 19 Apr 2011 10:02:52 +0000 (18:02 +0800)]
LU-198 Exit with error if NFSCLIENT is set, but no nfs export found
commit
2e7eb3a092fcc19a4550a677609f65f1fdd1e0d5
Date: Thu Jan 27 14:58:48 2011 +0100
b=24410 exit with error if NFSCLIENT is set, but no nfs export found
a=Elena Gryaznova
Signed-off-by: YangSheng <ys@whamcloud.com>
Change-Id: I8d25737fa1ebd19729230bd38b423aa028527f3a
Reviewed-on: http://review.whamcloud.com/409
Reviewed-by: Brian J. Murrell <brian@whamcloud.com>
Tested-by: Hudson
Reviewed-by: Yu Jian <yujian@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
yangsheng [Thu, 7 Apr 2011 13:41:29 +0000 (21:41 +0800)]
LU-197 Remove files after fsx in sanity-benchmark.sh to drop locks.
commit
9f1f7c4040440b8bab6130d30e9ecd1e8b803657
Date: Mon Jan 31 14:32:08 2011 +0100
b=23729 cancel_lru_locks not working cause some locks are still in cache from mmap files
Fix sanity-benchmark.sh to remove files after fsx otherwise client keeps locks acquired for mmap files in cache.
a=Johann Lombardi
Signed-off-by: YangSheng <ys@whamcloud.com>
Change-Id: Ib9683b5e8fce550259a1c3bfd617aa04c2f3abdc
Reviewed-on: http://review.whamcloud.com/408
Tested-by: Hudson
Reviewed-by: Yu Jian <yujian@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Brian J. Murrell [Tue, 3 May 2011 20:15:18 +0000 (16:15 -0400)]
LU-257 s/%krequires/%kversion/g
%krequires is not in the lustre.spec.in anymore after it was
removed due to being redundant (and unused) after bug 22281 was
landed.
Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: If27667f66e4b4d44845fb70d375699fead7da7e0
Reviewed-on: http://review.whamcloud.com/489
Reviewed-by: Richard Henwood <rhenwood@whamcloud.com>
Tested-by: Richard Henwood <rhenwood@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Hudson
Jinshan Xiong [Wed, 6 Apr 2011 00:24:53 +0000 (17:24 -0700)]
LU-185 LBUG: (cl_page.c:1362:cl_page_completion()) ...
!(pg->cp_flags & CPF_READ_COMPLETED) ASSERTION(0) failed
The root cause of this issue is that the application is doing regular and
direct IO on the same time. This will cause a ra page is issued for read
again.
Change-Id: I11bb46d334db52fa3faf3922b70ac831391a8217
Signed-off-by: Jinshan Xiong <jay@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/404
Tested-by: Hudson
Reviewed-by: wangdi <di.wang@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Johann Lombardi [Thu, 7 Apr 2011 18:55:29 +0000 (11:55 -0700)]
LU-130 disable page migration
Disable page migration until we implement a proper ->migratepage handler.
Change-Id: Ia98397db2ed6f6149b2708805320a5e929734308
Signed-off-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/399
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Reviewed-by: Lai Siyao <laisiyao@whamcloud.com>
Oleg Drokin [Sat, 23 Apr 2011 03:15:37 +0000 (23:15 -0400)]
LU-104 Properly address ownership of posix and flock locks
This patch fixes the long standing problem where we only relied
on pid to determine lock ownership.
This patch uses kernel-passed fl_owner field for posix lock and
fl_file for flock ownership like it should. This is a wire
protocol change of sorts, but I have compat code in place. Also
introduced a clear separation for on the wire and in-memory
policy representations and ways to convert between them.
This makes nfs4 locking work, also fixes original problem with
flock locking of the same fd from different processes. Makes
our own tests/flock_test.c to pass now which it never did before.
connectathon on nfs4 also passes locally for me now.
Added a flock test that is actually being run now.
Signed-off-by: Oleg Drokin <green@whamcloud.com>
Change-Id: Ifb4f4a0246f34143497225f2b8974f5fd93817a1
Reviewed-on: http://review.whamcloud.com/459
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Tested-by: Hudson
Niu Yawei [Tue, 19 Apr 2011 04:12:09 +0000 (21:12 -0700)]
LU-202 fix the test_23 of sanity-quota.sh
test_23() uses a constant value(120000 = ~120M) to check if the
system has enough space to finish the test, so in the large OSTCOUNT
testing environment, the test will fail for short of disk space.
Changes of this patch are:
- Use a reasonable value to check free space instead of magic 120000;
- Specify the stripe index of test file for test_9 and test_23;
Issue:
Signed-off-by: Niu Yawei <niu@whamcloud.com>
Change-Id: If7cd01b7ce9dc025a0f1b8b70fa9214d70182a86
Reviewed-on: http://review.whamcloud.com/434
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-by: Fan Yong <yong.fan@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Alexey Lyashkov [Sun, 3 Apr 2011 19:04:07 +0000 (23:04 +0400)]
LU-73 RHEL6 support.
Change-Id: If694f1e25cb19438bcc30cfa8fc42c01156c9f8a
Signed-off-by: Alexey Lyashkov <alexey_lyashkov@xyratex.com>
Signed-off-by: Vitaly Fertman <vitaly_fertman@xyratex.com>
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/394
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Jinshan Xiong [Fri, 1 Apr 2011 04:24:43 +0000 (21:24 -0700)]
LU-122 Revert the patch on bug 21122 and come up with a new fix
The patch on bug 21122 will cause deadlock. The root cause of
the issue is that a page was truncated even with cl_lock held.
Signed-off-by: Jinshan Xiong <jay@whamcloud.com>
Change-Id: If04a5632e64a019803465533d3ec7dba49e42168
Reviewed-on: http://review.whamcloud.com/316
Tested-by: Hudson
Reviewed-by: wangdi <di.wang@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Brian J. Murrell [Tue, 5 Apr 2011 05:16:51 +0000 (01:16 -0400)]
LU-175 don't allocate kernbuf on the stack
On i686, allocating a kernbuf of 1K on the stack causes a compiler
error:
lustre/mdt/mdt_lproc.c:255: error: the frame size of 1040 bytes is
larger than 1024 bytes
Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: I91e4fc39fa6bdc6b6d7937263520b1dfce8c4746
Reviewed-on: http://review.whamcloud.com/401
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Tested-by: Hudson
Reviewed-by: Lai Siyao <laisiyao@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Bobi Jam [Sat, 16 Apr 2011 04:09:52 +0000 (12:09 +0800)]
LU-216 Add WALK_SPACE_HAS_DATA_SEM check.
Change-Id: I4597ac7c22dec66859ba7f592beab968b1a04760
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/429
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Bobi Jam [Fri, 1 Apr 2011 04:15:17 +0000 (12:15 +0800)]
LU-172 Fix sanity check_seq_oid() for test_27z.
Change-Id: I7f839cf684d44289bcf73423602d9c128eed6cf8
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/389
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-by: Lai Siyao <laisiyao@whamcloud.com>
Niu Yawei [Fri, 1 Apr 2011 02:14:33 +0000 (19:14 -0700)]
LU-129 Skip quotacheck on administratively disabled OSTs
- skip quotacheck on administratively disabled OSTs.
- introduce OBD_NOTIFY_DEACTIVATE & OBD_NOTIFY_ACTIVATE event, thus lov is now aware of
administratively enable/disable import event, and is able to udpate the 'ltd_activate'
accordingly.
Issue: LU-129
Signed-off-by: Niu Yawei <niu@whamcloud.com>
Change-Id: I6af7dc09226d610bb1d481609a0a0ae09a05410e
Reviewed-on: http://review.whamcloud.com/387
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Yu Jian [Mon, 18 Apr 2011 08:38:39 +0000 (16:38 +0800)]
LU-203 pass client mount options to zconf_mount_clients() correctly
Quote the third argument and pass it to zconf_mount_clients().
Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: Ia29f2440894a9ec7a38615295ab1cd595d50c66f
Reviewed-on: http://review.whamcloud.com/430
Tested-by: Hudson
Reviewed-by: Chris Gearing <chris@whamcloud.com>
Reviewed-by: Brian J. Murrell <brian@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Alexey Lyashkov [Fri, 18 Mar 2011 20:13:45 +0000 (23:13 +0300)]
LU-147 avoid 8k obd device amount limit
increase number obd devices per host and
improve a lookup speed in device list via hashes.
Change-Id: I0357d0da42ce6fdd6e0846bb5d389df45f4db046
Signed-off-by: Alexey Lyashkov <Alexey_Lyashkov@xyratex.com>
Signed-off-by: Vitaly Fertman <vitaly_fertman@xyratex.com>
Reviewed-on: http://review.whamcloud.com/347
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
hongchao.zhang [Sat, 2 Apr 2011 06:26:33 +0000 (14:26 +0800)]
LU-212 reject connection with bad ocd_brw_size
in obdfilter, rejecting connection request with -EPROTO if OBD_CONNECT_BRW_SIZE
is set in obd_connect_data->ocd_connect_flags, but "ocd_brw_size" in it is zero.
Change-Id: Iba002250a11ab7321dbb5d4f398dde5ebd00c11d
Signed-off-by: Hongchao Zhang <hongchao.zhang@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/431
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Brian J. Murrell [Wed, 13 Apr 2011 13:59:28 +0000 (09:59 -0400)]
LU-210 backout debian packaging with patches
The debian packaging process included the [de-]applicaiton of patches
that were the commits made since the last tag. This provided greater
transparency as to exactly what was included in a release when it was
made between tags.
Unfortunately the process that was being used to achieve this failed
in a scenario where there were patches to files which were later removed
from the tarball making process.
For what it's worth, the right way to correct this and keep the
transparency is to have "make dist" create the tarball from the most
recent tag made and then have the patches generated that bring the
source up to current HEAD. In order to achieve this however both debian
and RPM packaging would need to operate in this manner. That would
mean adding Patch$n: lines to the RPM spec for each of the generated
patches. I'm not really sure we want to go there though.
Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: I9242799f3987e68e806e4c398e06cecbe1f5cc27
Reviewed-on: http://review.whamcloud.com/425
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Reviewed-by: Michael MacDonald <mjmac@whamcloud.com>
Liang Zhen [Mon, 4 Apr 2011 13:46:23 +0000 (21:46 +0800)]
LU-190 opencreate without mode can crash llite
user should specify mode for opencreate but if they don't llite will
LBUG, which is not good.
Change-Id: I614e655a8bef8ea9d62161591c99d305ce760732
Signed-off-by: Liang Zhen <liang@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/395
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
nasf [Sun, 27 Mar 2011 10:56:59 +0000 (18:56 +0800)]
LU-166 NOT assert wire data from client for getattr-by-fid
Change-Id: I44e96275ab53b8c7d301824487541b67b3c80cb8
Signed-off-by: nasf <yong.fan@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/367
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Reviewed-by: wangdi <di.wang@whamcloud.com>
Liang Zhen [Mon, 4 Apr 2011 16:31:20 +0000 (00:31 +0800)]
LU-191 mdd_close always set atime even it's same
mdd_close always set atime even new atime is exactly same as old one,
this will impact performance of those shortly open/close applications
and tests because of unnecessary transaction.
Change-Id: Ib408dd238fca2e8a736aff1874e243f0d26894b0
Signed-off-by: Liang Zhen <liang@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/396
Tested-by: Hudson
Reviewed-by: Mikhail Pershin <tappro@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Brian J. Murrell [Tue, 15 Mar 2011 00:33:33 +0000 (20:33 -0400)]
LU-127 RHEL packages should require the kernel
Since weak-modules is pretty much a non-starter on RHEL, we should
restore the Requires: kernel = $version dependency relationship between
the lustre modules and the kernel they were built for.
Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: I3f5d4501abd9751057cbc260b4ec3fb31ed18029
Reviewed-on: http://review.whamcloud.com/328
Tested-by: Hudson
Reviewed-by: Michael MacDonald <mjmac@whamcloud.com>
Reviewed-by: Yang Sheng <ys@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Bobi Jam [Mon, 11 Apr 2011 08:30:33 +0000 (16:30 +0800)]
LU-206 Memory corruption due to MMP (1.8.5)
Change-Id: Ia7c01488a7cc89193ff01a0086a34d42936844b5
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/420
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Alexey Lyashkov [Fri, 18 Mar 2011 20:11:40 +0000 (23:11 +0300)]
LU-147 hash code cleanup
Change-Id: I97cc182a9618271e59c4d6a8fbc2e887e1fb1a17
Signed-off-by: Alexey Lyashkov <Alexey_Lyashkov@xyratex.com>
Signed-off-by: Vitaly Fertman <vitaly_fertman@xyratex.com>
Reviewed-on: http://review.whamcloud.com/346
Tested-by: Hudson
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Liang Zhen <liang@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
hongchao.zhang [Wed, 23 Mar 2011 03:05:37 +0000 (11:05 +0800)]
LU-201 unregister bulk after unregistering reply
in ptlrpc_set_wait, the bulk is not unregistered if the reply was unregistered
asynchronously, which will cause "ptlrpc_set_wait" to run into dead loop
Change-Id: Ic79ae29c79be015269bb2b9bb80b5b4dd30be78d
Signed-off-by: Hongchao Zhang <hongchao.zhang@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/413
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Niu Yawei [Thu, 14 Apr 2011 02:50:37 +0000 (19:50 -0700)]
LU-184 Keep orphan on failover umount
- On failover umount, the orphan should not be cleared on last close.
- In mdt_mfd_close(), the ma_valid should not be set as zero, otherwise,
mds will never issue object destroy to OSS.
Issue:
Signed-off-by: Niu Yawei <niu@whamcloud.com>
Change-Id: If0a894655d0ea42ad0d07e725174b73a137b7adb
Reviewed-on: http://review.whamcloud.com/426
Tested-by: Hudson
Reviewed-by: Mikhail Pershin <tappro@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Christopher J. Morrone [Thu, 6 Jan 2011 01:07:49 +0000 (17:07 -0800)]
LU-25: Use LDLM_ASYNC with ldlm_cancel_lru to avoid blocking.
Have ldlm_cli_pool_shrink() call ldlm_cancel_lru() with the LDLM_ASYNC
option to have the blocking callback handled in another thread. This
will avoid randomly blocking for excessively long times when the shrinker
is called.
Change-Id: Ide705df2640a2a96be554ffc452930a2e180fc20
Signed-off-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-on: http://review.whamcloud.com/277
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Johann Lombardi [Sat, 9 Apr 2011 20:46:34 +0000 (13:46 -0700)]
LU-140 Remove leftovers of iopen
iopen has been already removed from the patch series, but it is still referenced in many places.
Change-Id: I512ce8ceef11d99f812f1348a5b53552b3ac1a42
Signed-off-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/417
Tested-by: Hudson
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Yu Jian [Wed, 6 Apr 2011 14:10:33 +0000 (22:10 +0800)]
LU-183 replace osd with osd* in proc paths
Fix the test scripts to use "osd*" instead of "osd" so that
it works with any OSD type.
Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: Ibc403ff146a7f429acb16c35449ec3026f6a3c7b
Reviewed-on: http://review.whamcloud.com/405
Tested-by: Hudson
Reviewed-by: Brian J. Murrell <brian@whamcloud.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Brian J. Murrell [Tue, 5 Apr 2011 02:49:34 +0000 (22:49 -0400)]
LU-73 add a kernel config for rhel6/i686
And a small patch to lbuild-rhel to display the diff of what got
built and what's in our tree for a .config.
Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: Ic847c92bbef5d6293cd54e26be242d37b5eb13c5
Reviewed-on: http://review.whamcloud.com/400
Tested-by: Hudson
Reviewed-by: Michael MacDonald <mjmac@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Brian J. Murrell [Thu, 3 Mar 2011 13:49:43 +0000 (08:49 -0500)]
LU-98 Fix defects in bug 14949 implementation
Some of the details of the implentation of bug 14949 "integrate
lustre-iokit into normal build process" were incorrect. This was leading
to a problem when trying to build with "make rpms" after configuring without
the iokit.
Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: Id064c688d63e3e70a90b725e49831b9c319229f0
Reviewed-on: http://review.whamcloud.com/294
Tested-by: Hudson
Reviewed-by: Robert Read <rread@whamcloud.com>
Reviewed-by: Michael MacDonald <mjmac@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Brian J. Murrell [Wed, 2 Mar 2011 18:10:55 +0000 (13:10 -0500)]
LU-115 get some verbose patching output
Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: I548a3a26b61241ed8d81551779bd78950dccdd54
Reviewed-on: http://review.whamcloud.com/298
Tested-by: Hudson
Reviewed-by: Robert Read <rread@whamcloud.com>
Reviewed-by: Michael MacDonald <mjmac@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Liang Zhen [Mon, 28 Mar 2011 16:05:16 +0000 (00:05 +0800)]
LU-170 oscc_grow_count will never grow
We are using req::rq_async_args.space[0] to store original value of
oscc_grow_count, and using req::rq_async_args.pointer_arg[0] to store
oscc, however, ptlrpc_async_args is a union, which means
req::rq_async_args.space[0] will always be overwritten by a ossc (pointer),
and osc_interpret_create will always get true on this condition
"if (diff < (int) req->rq_async_args.space[0])" and reset
oscc_grow_count to OST_MIN_PRECREATE and set OSCC_FLAG_LOW.
Because it's very unsafe to use raw scratchpad directly, I also cleaned
up all using of raw scratchpad in this patch.
Change-Id: I56348c2ebaf27acb493185db73f3992a17610d98
Signed-off-by: Liang Zhen <liang@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/371
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-by: Fan Yong <yong.fan@whamcloud.com>
Bobi Jam [Sun, 3 Apr 2011 04:14:38 +0000 (12:14 +0800)]
LU-189 RHEL6 MMP patch misses fix from bug 22117
Change-Id: I61010892c69f744e10c9561a202573fd08543d99
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/393
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Alexey Lyashkov [Fri, 18 Mar 2011 11:55:56 +0000 (14:55 +0300)]
LU-144 xattr fixes
- generate trusted.lov locally instead of ask MDT about it
- take conflicting LCK_EX lock to properly flush ALCs on update from client
Change-Id: Ic78f21fffb5174f83400e70ab368c80531322663
Signed-off-by: Alexey Lyashkov <Alexey_Lyashkov@xyratex.com>
Signed-off-by: Vitaly Fertman <vitaly_fertman@xyratex.com>
Reviewed-on: http://review.whamcloud.com/341
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Reviewed-by: Fan Yong <yong.fan@whamcloud.com>
ZhangHongChao [Tue, 31 Aug 2010 20:38:48 +0000 (13:38 -0700)]
LU-13 updated patch to fix at_min issue
Also Bugzilla bug 23352, attachment 31479
This is already landed for 1.8.6.
Change-Id: I87db1f40401e1da9337eace5d3bd992bc64871a4
Signed-off-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-on: http://review.whamcloud.com/306
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Reviewed-by: Mikhail Pershin <tappro@whamcloud.com>
Bobi Jam [Fri, 1 Apr 2011 01:37:18 +0000 (09:37 +0800)]
LU-182 Add ldiskfs 64bit-name-hash patch.
Change-Id: Ic00e17dbba630c8b1367ddfb6f6c350111b045b5
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/388
Tested-by: Hudson
Reviewed-by: Andrew Perepechko <andrew_perepechko@xyratex.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Bobi Jam [Thu, 31 Mar 2011 07:55:23 +0000 (15:55 +0800)]
LU-22 ldlm_resource::lr_lvb_data is protected by wrong lock
Use lr_lock to protect lr_lvbo_data while lr_lvbo_sem is only for
lvbo_init serialization.
Bugzilla: 24336
Change-Id: Ic3e77e99d8b3a3ca277adbc6548c254969e9761a
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/379
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Bobi Jam [Thu, 31 Mar 2011 04:06:14 +0000 (12:06 +0800)]
LU-73 Make register_shrinker argument more visible.
Change-Id: I0e1e9331b250bb31ae1e545e5705243e3a26c75a
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/377
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Reviewed-by: wangdi <di.wang@whamcloud.com>
nasf [Tue, 29 Mar 2011 06:53:38 +0000 (14:53 +0800)]
LU-168 Claim MDS_INODELOCK_LOOKUP lock when revalidate root object
Change-Id: I6854627ad1a168817d7551aee880652c873a663a
Signed-off-by: nasf <yong.fan@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/370
Tested-by: Hudson
Reviewed-by: Mikhail Pershin <tappro@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Brian J. Murrell [Fri, 25 Mar 2011 04:48:57 +0000 (00:48 -0400)]
LU-177 lbuild autonomy fixes
Fixes to allow lbuild to work without uneeded help from it's caller:
- kernel srpm location for rhel5
- CentOS is a "rhel" distro
Direct output to the output FD.
Refactor find_linux_rpm-rhel{5,6}.
Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: I714b90b53b14ac4ef4cd6bfe638a029656a151b0
Reviewed-on: http://review.whamcloud.com/361
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Tested-by: Hudson
Reviewed-by: Michael MacDonald <mjmac@whamcloud.com>
Johann Lombardi [Thu, 24 Mar 2011 17:01:04 +0000 (18:01 +0100)]
LU-155 cleanup some dead code
- 2.x does not support any kernels with vfs intents, so we can drop the related lustre code
(as a side note, HAVE_VFS_INTENT_PATCHES seems to be gone already)
- remove ll_pin/unpin since it won't be used any more
- add a comment to ll_ddelete()
Change-Id: I3ed16d84c8618071ea66d9c2cc9d2810b23f1830
Signed-off-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/359
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Reviewed-by: Mikhail Pershin <tappro@whamcloud.com>
nasf [Tue, 29 Mar 2011 03:13:28 +0000 (11:13 +0800)]
LU-167 Return "-ENOENT" instead of "-ESTALE" when getattr against non-existent object
Change-Id: I17148a361e711c92f31f6a9318512267a866c724
Signed-off-by: nasf <yong.fan@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/369
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Reviewed-by: wangdi <di.wang@whamcloud.com>
Tested-by: Hudson
Ned Bass [Fri, 11 Feb 2011 20:12:19 +0000 (12:12 -0800)]
LU-84 Allow building against kernel devel headers with quotas
The quota format definitions are normally found in private kernel
headers. However, some sites build Lustre against kernel development
headers rather than than full kernel source. To accomodate this we add
the minimal set of required definitions to
lustre/include/lustre_quota.h which will take effect only if
quotaio_v*.h are not located at configure time. The upstream version
of the 64-bit quota format patch in RHEL6 adds QFMT_VFS_V1 to
include/linux/quota.h, so configure can check for that if the
quotaio_v*.h headers aren't found.
Signed-off-by: Ned Bass <bass6@llnl.gov>
Change-Id: Id627f21604f1e35f550adb972c7f0e0ca9edf115
Reviewed-on: http://review.whamcloud.com/295
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Reviewed-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
nasf [Fri, 25 Mar 2011 05:17:49 +0000 (13:17 +0800)]
ORNL-2 MDS only processes child lock for getattr-by-fid
Change-Id: I4fd46761702f13173f43c83e1d76ff4b0478e786
Signed-off-by: nasf <yong.fan@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/365
Reviewed-by: wangdi <di.wang@whamcloud.com>
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Bobi Jam [Thu, 24 Mar 2011 17:06:06 +0000 (01:06 +0800)]
LU-73 RHEL6 support.
Add kernel jbd2-commit-timer-no-jiffies-rounding-rhel6.patch.
Change-Id: Ic0f49962ecb72b974a3e5d6225c2995e481c6783
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/358
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Bobi Jam [Thu, 24 Mar 2011 16:57:31 +0000 (00:57 +0800)]
LU-162 Export ext4_discard_preallocations for ldiskfs
Change-Id: I75d5553293fe5c4e409557895c48d02c65180409
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/360
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Johann Lombardi [Mon, 28 Feb 2011 14:49:21 +0000 (15:49 +0100)]
LU-91 Fix quota format problem with RHEL6 and kernels >= 2.6.33
Since kernel commit
869835dfad3eb6f7d90c3255a24b084fea82f30d "quota: Improve
checking of quota file header", quota users now have to specify what
format they want to use and an error is returned if it does not match the
one used on disk.
Since we only support the 64-bit quota format with lustre 2.x, we should
just use QFMT_VFS_V1 instead of QFMT_VFS_V0 (which should still be used
for older kernels like RHEL5).
Change-Id: I9022073385c76dafbf698fbdf2df58de38535d88
Signed-off-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/268
Reviewed-by: Brian J. Murrell <brian@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Tested-by: Hudson
Bobi Jam [Tue, 15 Mar 2011 01:19:12 +0000 (09:19 +0800)]
LU-73 RHEL6 support.
Include client, ldiskfs, kernel patches.
Change-Id: Ice16b8bf40c2e37df9af9f399316917097e8ee8f
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/307
Tested-by: Hudson
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Reviewed-by: Brian J. Murrell <brian@whamcloud.com>
Johann Lombardi [Thu, 17 Mar 2011 21:17:49 +0000 (22:17 +0100)]
LU-139 move DCACHE_LUSTRE_INVALID to higher bit to prevent conflict with other dentry flags
2.6.38 comes with a bunch of new dentry flags and one of them
(i.e. DCACHE_OP_REVALIDATE) conflicts with our DCACHE_LUSTRE_INVALID.
Change-Id: I11de689058d12b07b2f00191f2d175d48a90ff8e
Signed-off-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/339
Tested-by: Hudson
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Bobi Jam [Tue, 22 Mar 2011 03:01:52 +0000 (11:01 +0800)]
LU-39 ASSERTION(atomic_read(&client_stat->nid_exp_ref_count) == 0)
In lprocfs_exp_setup(), we need release old stats in all cases.
Bugzilla: 23499
Change-Id: Ida40ffe82a12e213779e344da92f5e3e437f8c8b
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/350
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Brian J. Murrell [Wed, 9 Mar 2011 21:11:07 +0000 (16:11 -0500)]
LU-73 lbuild and friends for RHEL6
This adds the lbuild support needed to build RHEL6 servers.
Included is a re-org of the RHEL5 lbuild code to maximize code
reuse and minimze code replication.
The kernel config here is the stock config from an x86_64 RHEL6
kernel build. It might need some tuning for Lustre.
Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: If83ee846d6d84a456444fb54feba0ad7810a586a
Reviewed-on: http://review.whamcloud.com/332
Tested-by: Hudson
Reviewed-by: Michael MacDonald <mjmac@whamcloud.com>
Reviewed-by: Robert Read <rread@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Jinshan Xiong [Sat, 19 Mar 2011 19:44:03 +0000 (12:44 -0700)]
LU-148 ll_readpage has to unlock vmpage by any means
Unlock vmpage in case ll_cl_init fails.
Change-Id: Id7d92df20c7057b324f21067e5b1cd3311cf9f41
Signed-off-by: Jinshan Xiong <jay@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/349
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Reviewed-by: wangdi <di.wang@whamcloud.com>
nasf [Tue, 1 Mar 2011 09:17:29 +0000 (17:17 +0800)]
b=24226 SUID/SGID related processing
1) remove SUID/SGID when writes/truncates file.
2) keep SUID/SGID for normal chmod without file data changed.
Issue: LU-65
Change-Id: I664f16c9bace1b0c011abcc7e2d103432886350e
Signed-off-by: nasf <yong.fan@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/238
Tested-by: Hudson
Reviewed-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Johann Lombardi [Mon, 7 Mar 2011 16:55:23 +0000 (17:55 +0100)]
LU-114 use ll_lookup_one_len() instead of lookup_one_len() in sptlrpc_target_local_copy_conf() should lock the parent dir when doing lookup
Recent kernels print a warning when lookup_one_len() is called w/o the parent dir locked.
sptlrpc_target_local_copy_conf() should not call lookup_one_len() directly, but ll_lookup_one_len()
instead which locks the parent dir for you and also checks for bad inode.
Change-Id: I4d020c55382238bbbc6df6017f54b5b585f53c48
Signed-off-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/303
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
nasf [Tue, 1 Mar 2011 07:59:40 +0000 (15:59 +0800)]
b=20581 MDS returns full hash for readdir to decrease hash collision
1) MDS returns full hash (both major and minor, 64bit) for readdir to decrease
hash collision.
2) Synchronization control between traversing thread and statahead thread when
readdir.
Issues: LU-71
Change-Id: I043784d58607d474a501111c9690fdab89ce8a4a
Signed-off-by: nasf <yong.fan@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/281
Tested-by: Hudson
Reviewed-by: Cliff White <cliffw@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Oleg Drokin [Mon, 14 Mar 2011 22:46:17 +0000 (18:46 -0400)]
LU-68 Fix a race between lock cancel and write
A race in osc_lock_detach due to disconnect between reading
kms from lock resource and write updating kms in the resource
could lead to loss of the update.
Lock the object across entire kms access to avert the issue.
Change-Id: I4b15e611b91b7a4a6b69f5c9fa9b6bd0a1f8cac9
Signed-off-by: Oleg Drokin <green@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/327
Tested-by: Hudson
Reviewed-by: wangdi <di.wang@whamcloud.com>
Reviewed-by: Jinshan Xiong <jay@whamcloud.com>
Jinshan Xiong [Mon, 7 Mar 2011 18:56:17 +0000 (10:56 -0800)]
LU-82 Remove useless clio locks
A CLS_NEW toplock must be deleted also if the last sublock is being deleted.
Signed-off-by: Jinshan Xiong <jay@whamcloud.com>
Change-Id: I9831ab534ae222fe74ebaf34faafe00bf3d18cb7
Reviewed-on: http://review.whamcloud.com/305
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Oleg Drokin [Fri, 11 Mar 2011 01:30:17 +0000 (20:30 -0500)]
LU-83 Kernel BUG in ext4 if wrong mount option is given
Original attempt at landing a fix for this bug landed hopelessly
broken patch.
Replace the patch with proper code.
Change-Id: Iddc38c4254b1f981b10e272b671fad232e20fd70
Signed-off-by: Oleg Drokin <green@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/314
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Oleg Drokin [Fri, 11 Feb 2011 03:28:15 +0000 (22:28 -0500)]
b=24375 Fix lvb updating race in enqueue vs completion case
ldlm_enqueue_tail checked for lock mode and updated lvb without
taking appropriate locks.
Take res and lock around mode check and lvb updating.
Issue: LU-67
Change-Id: I8fb764313326b8bf8f813e7cf77c050dd68afe45
Reviewed-on: http://review.whamcloud.com/229
Tested-by: Hudson
Reviewed-by: Liang Zhen <liang@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Niu Yawei [Fri, 18 Feb 2011 03:46:53 +0000 (19:46 -0800)]
b=22980 improve obdfilter-survey performance on multi-core system
1) Add unlock_ioctl support.
2) Land OBD_ALLOC_LARGE() patch from b4424 to avoid vmalloc in ioctl path.
3) Minor fix in class_handle_ioctl() to avoid double free.
Issue: LU-66
Change-Id: Id965cd13fdaeb63197c30e557708afbfa9ee9f39
Signed-off-by: Niu Yawei <niu@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/247
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Yu Jian [Mon, 7 Mar 2011 10:44:02 +0000 (18:44 +0800)]
LU-116 fix mmp test 9 and test 10
1) fix the sleep time in test 9 to abort e2fsck before it's complete
2) fix test 10 to allow e2fsck run on mounted filesystem with -n option
Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: I2605d75f47e44db659ced2e5435e3b51cc8897c6
Reviewed-on: http://review.whamcloud.com/301
Tested-by: Hudson
Reviewed-by: Brian J. Murrell <brian@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
nasf [Wed, 2 Mar 2011 10:23:08 +0000 (18:23 +0800)]
b=22703 no error when close if application has known write failure already
Do not return error when close if application has known write failure already.
Issues: LU-65
Change-Id: I25b9628fcdd3bf5b7db9d3f8460e06c18226b2e0
Signed-off-by: nasf <yong.fan@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/286
Tested-by: Hudson
Reviewed-by: Jinshan Xiong <jay@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Christopher J. Morrone [Thu, 24 Feb 2011 22:11:45 +0000 (14:11 -0800)]
Make llite_lloop.ko build conditional
The llite_lloop.ko module only works with kernel page sizes smaller
than 64K, so test for the kernel page size and only build the module
conditionally.
Issue: LU-94
Change-Id: I78e416832767bbe29529d152e45addec7a999506
Signed-off-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-on: http://review.whamcloud.com/266
Tested-by: Hudson
Reviewed-by: Brian J. Murrell <brian@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
nasf [Mon, 7 Mar 2011 15:23:42 +0000 (23:23 +0800)]
ORNL-2 Enable OBD_CONNECT_ATTRFID supporting on lustre-2.x
To drop redundant MDS-side RPC for root object revalidation when
stats each item for traversing directory based on absolute path.
Signed-off-by: nasf <yong.fan@whamcloud.com>
Change-Id: I5f7c7755b4142ac1f23c66517e5919c1398e15b8
Reviewed-on: http://review.whamcloud.com/302
Tested-by: Hudson
Reviewed-by: Lai Siyao <laisiyao@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Brian J. Murrell [Fri, 21 Jan 2011 22:47:58 +0000 (17:47 -0500)]
b=24417 Update Build-Depends
- remove texlive-latex-recommended as a build requirement
- add missing "| automake1.7 | automake1.8 | automake1.9" to
debian/control.main
Issue: LU-52
Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: Ief22c678014836e8372a18dd3ed48a95631177b5
Reviewed-on: http://review.whamcloud.com/212
Tested-by: Hudson
Reviewed-by: Robert Read <rread@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Brian J. Murrell [Mon, 7 Mar 2011 22:37:44 +0000 (17:37 -0500)]
LU-120 the perfect storm
There is a corner case in the debian patching/unpatching process
where a patch converts the use of a given file to a file.in that is
to be substed. The problem is that between the configure and the
make debs, a file, which a debian patch wants to remove already
exists (it was created by the configure) and dpkg-buildpackage's
unpatch step fails becuase it wants to try to create a file which
configure already created.
Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: I185e496b82687d21ea3d744e939536c106b702f9
Reviewed-on: http://review.whamcloud.com/308
Tested-by: Hudson
Reviewed-by: Robert Read <rread@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Brian J. Murrell [Thu, 24 Feb 2011 18:06:24 +0000 (13:06 -0500)]
LU-97 lustre-modules should depend on ldiskfs.
Only for the server of course.
But depend on a generically named tag, and have ldiskfs provide
that tag so that any future backing stores can be used as a
replacement for ldiskfs.
Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: Ibef2b8c2b2a8c7339143dc32d32b59ace94f00fd
Reviewed-on: http://review.whamcloud.com/265
Reviewed-by: Robert Read <rread@whamcloud.com>
Tested-by: Hudson
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Robert Read [Tue, 1 Mar 2011 00:46:16 +0000 (16:46 -0800)]
LU-103 Add support for test groups.
Added an -g GROUP_FILE option to auster. The tests listed in the
group file will be used instead of any passed on the command line. I also added
a couple of test-groups to get things started, although
I'm sure we'll want to refine these.
Took care of a few cleanups in run_suite as well.
Signed-off-by: Robert Read <rread@whamcloud.com>
Change-Id: I5000c2b95a6eb1a49f5305637d4b55de5edc6d5c
Reviewed-on: http://review.whamcloud.com/278
Tested-by: Hudson
Reviewed-by: Chris Gearing <chris@whamcloud.com>
Reviewed-by: Wei Liu <sarah@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Brian J. Murrell [Mon, 7 Mar 2011 17:47:05 +0000 (12:47 -0500)]
LU-119 Remove code added by a conflict resolution
I erroneously re-added some code that was removed in
62345145fa7cae1846f96078bac43ddf58cea640 while resolving
a conflict in LU-85's landing of commit
ab13b69b63ca19d7fe68b8ba6afd2cb4723db78c.
Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: Id2489a3aa5560cdf4e7cf4755ddc16437558a2fc
Reviewed-on: http://review.whamcloud.com/304
Tested-by: Hudson
Reviewed-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-by: Robert Read <rread@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
James Simmons [Fri, 18 Feb 2011 18:58:41 +0000 (13:58 -0500)]
No bugzilla for this. One of the test failing was config-sanity.sh
test number 37. So I tracked it down to way. The reason was the mounts
were failing. Upon investigation I discovered the test caught a error
reported back by the mount. I tracked it down to the section of code in
the patch. Basically mount_lustre.c attempts to set the stripe_cache_size
for /dev/mdX devices. The problem is this extra feature is present for
certain classes of md devices. In the case of our test bed we are one
of the classes that lacks this enhancement. So instead of failing to mount
in that case we just print a warning that the special param is not present
but still succed to mount.
Change-Id: Ib8a743e77b5e6994e2640c51b781f7d0d345d5a7
Signed-off-by: James Simmons <uja.ornl@gmail.com>
Reviewed-on: http://review.whamcloud.com/254
Tested-by: Hudson
Reviewed-by: Brian J. Murrell <brian@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Brian J. Murrell [Wed, 23 Feb 2011 21:38:40 +0000 (16:38 -0500)]
LU-85 Update some URLs and text to reflect our releases.
The text and URLs in the debian packaging metadata were old
(circa Sun Microsystems), so update them and take the opportunity
to inform about the source of the release being installed.
Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: I7860e8ada7757d8a991a472af293749fd4c89f5a
git://git.whamcloud.com / fs / lustre-release.git / log