Whamcloud - gitweb
Vladimir Saveliev [Tue, 14 Dec 2010 12:30:12 +0000 (15:30 +0300)]
b=13698 llapi_get_version
this uses OBD_GET_VERSION ioctl to obtain lustre version
i=andreas.dilger
Elena Gryaznova [Mon, 13 Dec 2010 13:32:55 +0000 (16:32 +0300)]
b=23961 fix for setup with several network interfaces
i=Andrew.Perepechko
- metadata-updates fix for setup when several interfaces are UP on host;
hostname could be assigned to IP which is different from lnet network used,
the hostname-s of NODES_TO_USE are now stored in HOSTS
- new SHUTDOWN_ATTEMPTS: the tunable number of attepts to shutdown node
- shutdown_node_hard () fix: do not call "power off" each time,
wait that the node is not pingable before the next "power off" attempt
- unused check_port() is removed
root [Fri, 10 Dec 2010 09:34:05 +0000 (02:34 -0700)]
b=4424 Reserve obd_connect_data.ocd_max_easize field
To avoid potential incompatible changes between b1_8 and master,
reserve the ocd_max_easize field. The corresponding connect flag
OBD_CONNECT_MAX_EASIZE has been reserved for some time already.
Add several other OBD_CONNECT_ flags that have already been defined
to the wirecheck/wiretest tools.
Elena Gryaznova [Wed, 8 Dec 2010 20:32:43 +0000 (23:32 +0300)]
b=22376 sanity test for non-root exec-only file execution
o=Andreas.Dilger
i=grev
Andrew Perepechko [Thu, 2 Dec 2010 23:47:38 +0000 (02:47 +0300)]
b=23766 interop bits for sanity/203
i=Elena Gryaznova
Elena Gryaznova [Thu, 2 Dec 2010 16:27:50 +0000 (19:27 +0300)]
b=24118 test_70b rundbench load failed
i=Brian.Murrell
i=Jian.Yu
- give rundbench a chance to start before the dbench load check
- new check_for_process () and killall_process () to check/kill any defined progs
instead of "dbench" only
- fix 70a, 70b to mount the clients on MOUNT instead of DIR
Andrew Perepechko [Wed, 1 Dec 2010 22:49:02 +0000 (01:49 +0300)]
Revert "b=23913 fix "ASSERTION(!cfs_list_empty(&dquot->dq_hash)) failed""
This reverts commit
9491d5d931407ccf68d95932ca4ea3c545c6f86d.
Elena Gryaznova [Wed, 1 Dec 2010 17:25:24 +0000 (20:25 +0300)]
b=24228 fix test duration check to be more accurate
i=vitaly
Elena Gryaznova [Tue, 30 Nov 2010 16:20:43 +0000 (19:20 +0300)]
b=23535 sgpdd-survey.sh should check for sg_map
i=Brian.Murrell
check that iokit sgpdd-survey and sg_map are installed
Elena Gryaznova [Tue, 30 Nov 2010 14:34:46 +0000 (17:34 +0300)]
b=22157 combined mgs/mds fix for single node setup
i=Mikhail.Pershin
for configuration combined mgs/mds on single node setup
we do not need to unload the modules because conf-sanity keeps
the mgs mounted during all tests
Elena Gryaznova [Tue, 30 Nov 2010 12:51:40 +0000 (15:51 +0300)]
b=23402 mmp_fini () multiple oss fix
i=Yu.Jian
Isaac Huang [Mon, 29 Nov 2010 04:22:42 +0000 (21:22 -0700)]
b=23575 O2iblnd credit deadlock regression
This fixed a regression of bug 14425.
i=liang
Landen [Mon, 29 Nov 2010 08:07:11 +0000 (16:07 +0800)]
b=23913 fix "ASSERTION(!cfs_list_empty(&dquot->dq_hash)) failed"
1. In an error handling, free_dquot() should be used.
2. Check if quota is on in dqacq_handler()
i=johann
i=panda
Landen [Mon, 29 Nov 2010 08:07:12 +0000 (16:07 +0800)]
b=23868 fix "sanity-quota test_18c: @@@@@@ FAIL: quotaon failed!"
i=johann
i=mikhail.pershin
Elena Gryaznova [Fri, 26 Nov 2010 18:18:36 +0000 (21:18 +0300)]
b=23954 MGS device has stopped when we try to start the second mgs
i=Mikhail.Pershin
add test_24b to ALWAYS_EXCEPT list for configuration mgs/mds are not combined
Elena Gryaznova [Fri, 26 Nov 2010 17:43:09 +0000 (20:43 +0300)]
b=23869 HARD failure mode fixes
i=Brian.Murrel
i=Wei.G.Li
facet_failover() has to restart only those affected facets which were UP
before the node failure
replay-single tests which use shutdown_facet() && reboot_facet() instead of
facet_failover() have to take care about the affected facets
Elena Gryaznova [Fri, 26 Nov 2010 17:12:33 +0000 (20:12 +0300)]
b=23956 change conf-sanity test_37 to be functional on remote setup
i=Mikhail.Pershin
fix test_37 to not be skipped on remote setup;
use the existing mds device instead of create a new one
Elena Gryaznova [Fri, 26 Nov 2010 16:36:06 +0000 (19:36 +0300)]
b=24020 lustre doesn't start with ext4 based ldiskfs.
i=Andrew.Perepechko
Andrew Perepechko [Fri, 26 Nov 2010 01:29:34 +0000 (04:29 +0300)]
b=24201 add procfs tunable to enable/disable lockless direct I/O
llite.lustre-*.lockless_direct_io=0 will disable default semantics
of direct I/O that forces it to be lockless. lockless_direct_io value,
however, will be ignored if per-file LL_FILE_LOCKED_DIRECTIO bit
is set.
i=Johann Lombardi
Johann Lombardi [Thu, 25 Nov 2010 17:36:38 +0000 (18:36 +0100)]
b=21804 make sure the request is protected by rq_refcount while
i=vitaly
i=oleg
Johann Lombardi [Thu, 25 Nov 2010 17:22:45 +0000 (18:22 +0100)]
b=23820 add missing changelog entry
Johann Lombardi [Thu, 25 Nov 2010 17:21:16 +0000 (18:21 +0100)]
b=21760 start bulk unregistering at the same time as reply unlink
i=vitaly
i=oleg
Dmitry Zogin [Sat, 2 Oct 2010 03:33:18 +0000 (23:33 -0400)]
b=23820 ptlrpc_check_set()) ASSERTION(req->rq_phase == RQ_PHASE_BULK) failed
i=oleg.drokin
i=johann.lombardi
Handle unsent requests with rq_net_err in ptlrpc_check_set().
Johann Lombardi [Thu, 25 Nov 2010 15:56:42 +0000 (16:56 +0100)]
move to 1.8.6
Andrew Perepechko [Thu, 18 Nov 2010 12:40:18 +0000 (15:40 +0300)]
fixed a typo from previous commit
Andrew Perepechko [Thu, 18 Nov 2010 12:35:02 +0000 (15:35 +0300)]
b=23766 atime is not updated properly
a test case for atime updates on an MDS
i=Johann Lombardi
i=Andreas Dilger
yangsheng [Mon, 15 Nov 2010 14:43:46 +0000 (22:43 +0800)]
b=23985 Fix upstream jbd2 code issue.
The jbd2_stats_proc_init() was placed on wrong location in
jbd2_journal_init_dev(). This may cause /proc/fs/jdb2/<dev>/*
cannot be created when using external journal device.
i=adilger
i=johann
Andrew Perepechko [Mon, 15 Nov 2010 12:43:06 +0000 (15:43 +0300)]
Changelog update
Andrew Perepechko [Mon, 15 Nov 2010 12:29:06 +0000 (15:29 +0300)]
b=24051 sanity test_76 fix
Due to per-cpu slab caches, "active" stat from /proc/slabinfo
is not always accurate.
i=ZhiYong Tian
Andrew Perepechko [Mon, 15 Nov 2010 12:21:11 +0000 (15:21 +0300)]
b=23766 atime isn't properly updated on the MDS
Based on the patch from Robin Humble (ANU)
i=Andreas Dilger
i=Vitaly Fertman
Andrew Perepechko [Wed, 27 Oct 2010 04:54:18 +0000 (12:54 +0800)]
b=21587 don't LBUG if transno has changed during replay
i=johann
i=tappro
yangsheng [Fri, 22 Oct 2010 09:43:38 +0000 (17:43 +0800)]
b=23988 Changes sd-iostats patch for 2.6.32 kernel.
Also statistic with REQ_TYPE_BLOCK_PC command.
Johann Lombardi [Thu, 21 Oct 2010 12:27:41 +0000 (14:27 +0200)]
Revert "b=23988 Changes sd-iostats patch for 2.6.32 kernel."
This reverts commit
50a915fdbb340f04a850778870d027c4c3303aab.
yangsheng [Thu, 21 Oct 2010 12:06:52 +0000 (20:06 +0800)]
b=23965 Add fiemap patch for ldiskfs base on ext4(sles11 sp1).
i=girish
i=johann
yangsheng [Thu, 21 Oct 2010 12:00:24 +0000 (20:00 +0800)]
b=23988 Changes sd-iostats patch for 2.6.32 kernel.
yangsheng [Thu, 21 Oct 2010 10:42:00 +0000 (18:42 +0800)]
b=24003 Shows stripe_count: 65535 but should be '-1'.
i=johann
Johann Lombardi [Wed, 20 Oct 2010 17:46:23 +0000 (19:46 +0200)]
b=21587 don't clean up echo_client is refcount is not 0
i=adilger
Johann Lombardi [Wed, 20 Oct 2010 09:37:53 +0000 (11:37 +0200)]
b=23966 Revert "b=21137 sles11 with 1.8 is slower than 1.6 sles10 for O_DIRECT single file IOR writes"
This reverts commit
1242fac546a8c252dee5715bb218133795211028.
This introduces issues with SATA drivers. See bug 23966.
Brian J. Murrell [Thu, 14 Oct 2010 18:34:15 +0000 (20:34 +0200)]
b=23721 download OFED from OFA
Download the OFED distribution directly from OFA to ameliorate
re-distribution complications.
i=yangsheng
i=wangyb
Johann Lombardi [Thu, 14 Oct 2010 16:42:27 +0000 (18:42 +0200)]
fix nit in lbuild-sles10
Terry Rutledge [Thu, 14 Oct 2010 00:11:32 +0000 (17:11 -0700)]
Updated for Lustre 1.8.5.
Johann Lombardi [Wed, 13 Oct 2010 22:41:11 +0000 (00:41 +0200)]
fix changelog
Johann Lombardi [Wed, 13 Oct 2010 22:35:22 +0000 (00:35 +0200)]
set version to 1.8.5.0 for RC1
yangsheng [Wed, 13 Oct 2010 14:22:24 +0000 (22:22 +0800)]
b=22514 Update to RHEL5.5 & OEL5.5 latest kernel.
Johann Lombardi [Wed, 13 Oct 2010 21:53:05 +0000 (23:53 +0200)]
b=20744 add changelog entry and set REBUILD_OFED_DEVEL_RPM for patched kernel only
Johann Lombardi [Wed, 13 Oct 2010 09:34:26 +0000 (11:34 +0200)]
b=23911 enable REBUILD_OFED_DEVEL_RPM by default
Brian J. Murrell [Tue, 12 Oct 2010 19:56:32 +0000 (15:56 -0400)]
b=20744 set arch correctly for ofed-devel
Oh boy, more fallout from our not building i586 kernels for sles.
If we are building lustre servers and NOT rebuilding the Suse ofed
RPM, then we need to look for the i586 ofed-devel package.
i=yangsheng
Johann Lombardi [Tue, 12 Oct 2010 15:55:46 +0000 (17:55 +0200)]
b=20744 add changelog entry
yangsheng [Tue, 12 Oct 2010 15:50:36 +0000 (23:50 +0800)]
b=20744 Update SLES10 SP3 kernel 2.6.16.60-0.69.1.
i=johann
i=brian
Dmitry Zogin [Tue, 12 Oct 2010 14:38:03 +0000 (10:38 -0400)]
b=21137 sles11 with 1.8 is slower than 1.6 sles10 for O_DIRECT single file IOR writes
Fix __bio_add_page() in order to prevent IO split on some drivers.
i=andreas.dilger
i=sheng.yang
Andrew Perepechko [Tue, 12 Oct 2010 12:33:43 +0000 (16:33 +0400)]
b=20533 set the default max_sectors to the raid5/6 stripe size
a=Kevin Van Maren
i=Jinshan Xiong
i=Atul Vidwansa
Andrew Perepechko [Tue, 12 Oct 2010 08:45:35 +0000 (12:45 +0400)]
b=23701 reduce stack pressure by uninlining some ptlrpc and mds routines
i=Oleg Drokin
i=Alexander Zarochentsev
i=Johann Lombardi
Elena Gryaznova [Mon, 11 Oct 2010 14:17:59 +0000 (18:17 +0400)]
b=23542 recovery-small: test_27 is not time bounded
i=Brian.Murrell
Elena Gryaznova [Mon, 11 Oct 2010 16:35:37 +0000 (20:35 +0400)]
b=23537 do_rpc_nodes does not propagate NAME to remote nodes causing rpc.sh to fail abruptly
o=Chris.Horn <hornc@cray.com>
i=grev
Johann Lombardi [Mon, 11 Oct 2010 13:11:23 +0000 (15:11 +0200)]
bump version to 1.8.4.55
Johann Lombardi [Mon, 11 Oct 2010 12:12:00 +0000 (14:12 +0200)]
b=23781 fix another potential obdo leak
i=hongchao
Brian Murell [Mon, 11 Oct 2010 12:02:16 +0000 (14:02 +0200)]
b=20744 remedial fix for inkernel ofed handling
i=yangsheng
It seems that the patch to add inkernel handling for SLES10's OFED introduced a
bug for other inkernel uses. This fixes that bug.
Johann Lombardi [Fri, 8 Oct 2010 16:16:37 +0000 (18:16 +0200)]
bump version to 1.8.4.54
Vladimir Saveliev [Fri, 8 Oct 2010 13:39:06 +0000 (17:39 +0400)]
b=23588 wait for new OST addition propagation to clients
i=nathan.rutman
Brian J. Murrell [Fri, 8 Oct 2010 14:17:41 +0000 (10:17 -0400)]
b=20744 hacks for Novell bug 639581
Brian J. Murrell [Fri, 8 Oct 2010 14:17:40 +0000 (10:17 -0400)]
b=20744 hacks for Novell bug 578796
These hacks are needed to deal with Novell bug 578796.
Brian J. Murrell [Fri, 8 Oct 2010 14:17:39 +0000 (10:17 -0400)]
b=20744 use Novell supplied ofed-devel
Several changes to support using ofed-devel supplied in SLES10 SP3
(currently as PTFs).
Fix a bug with passing arguments to rpm2cpio that should not be passed.
Add a hook for distro specific module building after the kernel[-devel]
build and unpack.
Fix resolve_arch to be able to resolve either the "canonical" arch for
i?86, or the arch that the patchless kernel is expected to be.
Bail out of download_ofed earlier if using "inkernel".
Fix a bug where an rpm2cpio command was not being error checked.
Fix a bug where we were not processing and making available files
in all of the RPMS/i?86 subdirs.
i=mjmac
i=wangyb
Brian J. Murrell [Wed, 6 Oct 2010 21:13:37 +0000 (17:13 -0400)]
b=23685 don't define unused variables
When CONFIG_SYSCTL is not defined in the kernel config, a couple of
variables are left unused, yet still defined.
i=Dmitry Zogin
i=Andrew Perepechko
Brian J. Murrell [Wed, 6 Oct 2010 21:13:36 +0000 (17:13 -0400)]
b=23804 LC_WALK_SPACE_HAS_DATA_SEM on servers only
The LC_WALK_SPACE_HAS_DATA_SEM check is only needed on servers
so guard it for server configures only.
i=Sheng Yang
i=Andreas Dilger
Andrew Perepechko [Fri, 8 Oct 2010 12:38:13 +0000 (16:38 +0400)]
b=23721 disable rdac and mptlinux by default
Disable the (downloading and) building of the MPTLINUX and RDAC
packages by default and provide a commandline switch to enable them.
i=yangsheng
i=wangyb
yangsheng [Fri, 8 Oct 2010 09:43:15 +0000 (17:43 +0800)]
b=21610 Update to sles11 sp1 latest kernel 2.6.32.19-0.2.
i=adilger
i=girish
Johann Lombardi [Fri, 8 Oct 2010 09:19:48 +0000 (11:19 +0200)]
Add missing changelog entries
Johann Lombardi [Thu, 7 Oct 2010 22:25:32 +0000 (00:25 +0200)]
Revert "b=21760 Application hung in direct I/O"
This reverts commit
91d48dd54d07c3901da34227037a32412243e08a.
With this patch, ptlrpcd sleeps until bulk buffers are unlinked by
the network stack. Meanwhile, no requests/replies are processed
by ptlrpcd which can cause various kind of problems (e.g. eviction).
Johann Lombardi [Thu, 7 Oct 2010 20:05:25 +0000 (22:05 +0200)]
b=23645 remove another i_mutex warning with 2.6.32
hold inode mutex before calling lookup_on_len() when creating
objects on MDS.
Andrew Perepechko [Thu, 7 Oct 2010 14:27:39 +0000 (18:27 +0400)]
b=23596 account direct i/o inflight separately from non-direct i/o
Account direct i/o inflight rpcs separately from non-direct i/o so
that direct i/o, which is limited by max_rpcs_in_flight, should not
block non-direct i/o, which is not limited by max_rpcs_in_flight.
i=Oleg Drokin
i=Alexander Zarochentsev
i=Johann Lombardi (author of the original patch)
Johann Lombardi [Thu, 7 Oct 2010 08:06:12 +0000 (10:06 +0200)]
b=23878 add debug info
Johann Lombardi [Thu, 7 Oct 2010 07:55:36 +0000 (09:55 +0200)]
Add support for OFED 1.5.2
Eric Mei [Wed, 6 Oct 2010 15:16:23 +0000 (09:16 -0600)]
b=22168 write_disjoint: retry if write receives EINTR.
o=robert.read
r=eric.mei
Andrew Perepechko [Wed, 6 Oct 2010 15:17:45 +0000 (19:17 +0400)]
b=23216 a fix for a possible memory leak in echo_client_kbrw
b=Oleg Drokin
b=Hong-Chao Zhang
Andrew Perepechko [Wed, 6 Oct 2010 15:14:25 +0000 (19:14 +0400)]
Merge branch 'b1_8' of git.lustre.org:prime/lustre into b1_8
Andrew Perepechko [Wed, 6 Oct 2010 15:12:52 +0000 (19:12 +0400)]
b=23827 fix per-NID reporting on outstanding writes
a=Kevin Van Maren
i=Johann Lombardi
i=Andreas Dilger
Johann Lombardi [Wed, 6 Oct 2010 13:51:13 +0000 (15:51 +0200)]
b=23780 hold write reference on i_data_sem for walk_space
i=girish
Restore the old behavior and live with the fiemap locking issue for now.
Andrew Perepechko [Tue, 5 Oct 2010 17:55:38 +0000 (21:55 +0400)]
b=23701 a build fix
Andrew Perepechko [Tue, 5 Oct 2010 16:56:14 +0000 (20:56 +0400)]
b=23710 reduce stack pressure by uninlining mds_log_lost_precreated
i=Johann Lombardi
Elena Gryaznova [Tue, 5 Oct 2010 16:46:33 +0000 (20:46 +0400)]
b=22546 sanity-quota test_18c: @@@@@@ FAIL: expect
104857600, got
102760448. Verifying file failed!
o=James.Simmons <simmonsja@ornl.gov>
i=Brian.Murrell
i=grev
i=landen
Andrew Perepechko [Tue, 5 Oct 2010 16:43:18 +0000 (20:43 +0400)]
b=23701 reduce stack pressure in ptlrpc_server_handle_request
i=Oleg Drokin
i=Alexander Zarochentsev
i=Johann Lombardi
Terry Rutledge [Fri, 1 Oct 2010 14:53:48 +0000 (07:53 -0700)]
Updated for 1.8.5 build 03.
Andrew Perepechko [Thu, 30 Sep 2010 12:45:54 +0000 (16:45 +0400)]
b=22107 a fix for the warning message in filter_destroy_internal
filter_destroy_internal is entered with inode references from
lookup and from the resource used by the extent lock, so the
check is changed to i_count != 2.
i=Vitaly Fertman
i=Johann Lombardi
Vladimir Saveliev [Wed, 29 Sep 2010 21:21:37 +0000 (01:21 +0400)]
b=22766 cascading_rw: take lmm_stripe_count returned by ioctl(LL_IOC_LOV_SETSTRIPE)
+cleanup in ll_dir_ioctl(): do not save return code of copy_to_user and copy_from_user
i=adilger
yangsheng [Mon, 20 Sep 2010 14:32:01 +0000 (22:32 +0800)]
b=22770 Removed the LASSERT since (conn == NULL) is a legitimate case.
i=adigler
i=panda
Johann Lombardi [Tue, 28 Sep 2010 14:01:37 +0000 (16:01 +0200)]
bump version to 1.8.4.52
hongchao.zhang [Sun, 12 Sep 2010 19:52:59 +0000 (03:52 +0800)]
b=23781 fix obdo leak issue
in ll_setattr_raw, if ll_extent_lock call fails, the newly allocated
obdo has not been freed before return.
i=johann
Dmitry Zogin [Fri, 24 Sep 2010 19:05:47 +0000 (15:05 -0400)]
b=23706 sles11sp1: jbd2 stat patch is missing
Temp patch to disable obdfilter-survey.sh 1b/2b tests,
if jbd stats are not available.
i=grev
Brian J. Murrell [Thu, 23 Sep 2010 15:23:34 +0000 (11:23 -0400)]
explicitly set KVERS; allow all automake packages
Explicitly set KVERS to what it'd default value would be if it's
not set by the caller.
Allow any of the compatible automake* packages in debian/control.
Girish Shilamkar [Thu, 2 Sep 2010 18:03:34 +0000 (23:33 +0530)]
b=22117 Patch to limit mmp interval.
i=adilger
i=kalpak
Brian J. Murrell [Sun, 19 Sep 2010 04:27:13 +0000 (00:27 -0400)]
move .debs into /debs
Move the built .debs from .. into a /debs subdir of the build tree.
We in fact could almost make debian/rules do this using dh_builddeb's
--destdir argument, however dpkg-genchanges wants to find them in ..
and there is no way to pass dpkg-genchange's -u argument to it from
dpkg-buildpackage. So thus, our hack.
Elena Gryaznova [Thu, 16 Sep 2010 18:40:22 +0000 (22:40 +0400)]
b=22934 conf_sanity writeconf method is never tested
i=Andrew.Perepechko
use the correct mds dev
Dmitry Zogin [Thu, 16 Sep 2010 12:16:41 +0000 (08:16 -0400)]
b=17682 Improve debug capability for production systems
Improve performance of D_RPCTRACE debugging.
i=andreas.dilger
i=andrew.perepechko
Dmitry Zogin [Thu, 16 Sep 2010 12:13:08 +0000 (08:13 -0400)]
b=22378 Correct MDS client stats
Interoperability fix for test_133[abc] sanity.sh tests
o=grev
i=dmitry.zoguine
Andreas Dilger [Wed, 1 Sep 2010 09:07:53 +0000 (03:07 -0600)]
Ignore generated ll_decode_filter_fid file
Vladimir Saveliev [Tue, 14 Sep 2010 11:45:05 +0000 (15:45 +0400)]
b=23271 handle e2fsck and lfsck return values
to avoid test-framework.sh' trap
i=grev
i=jian.yu
Andrew Perepechko [Thu, 16 Sep 2010 11:12:40 +0000 (15:12 +0400)]
b=20101 lfs ost enhancements
a=James Simmons <simmonsja@ornl.gov>
i=Andreas Dilger
i=Brian Murrell
Rahul Deshmukh [Thu, 16 Sep 2010 08:08:09 +0000 (13:38 +0530)]
b=23200 1.8<->2.0 interop: sanityN test_12: @@@@@@ FAIL: test_12 failed with 1
b=23200 1.8<->2.0 interop: sanityN test_12: @@@@@@ FAIL: test_12 failed with 1
This patch fixed the 1.8<->2.0 interop: sanityN test_12
i=rread
Andrew Perepechko [Wed, 15 Sep 2010 21:11:56 +0000 (01:11 +0400)]
b=23216 a fix for a possible memory leak in filter_server_init_data
i=Alexander Zarochentsev
i=Dmitry Zogin
Dmitry Zogin [Wed, 15 Sep 2010 19:41:32 +0000 (15:41 -0400)]
b=21137 sles11 with 1.8 is slower than 1.6 sles10 for O_DIRECT single file IOR writes
Extra debug logging in filter_do_bio()
i=andreas.dilger
i=andrew.perepechko