Whamcloud - gitweb
Andreas Dilger [Thu, 26 Jan 2012 12:39:06 +0000 (05:39 -0700)]
LU-797 tests: speed up ost-pools tests
The test time of the ost-pools subtests is unreasonably long.
test_14 fills an OST to 90% full, regardless of the OST size.
Skip the test if the amount of data to be written is too large
to run in a practical time.
test_18 creates 3x3x30000 files to compare performance with/without
pools enabled. Instead of creating a fixed number of files, use
createmany to run for a specific (short) time to measure performance.
test_23 tried to fill all OSTs 100% full. Split this test into two:
- test_23a to test quota with a file in a pool
- test_23b to test OOS with a file striped over pool
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: Ib851939a210ab68f52da1ac777781c2a922c500c
Reviewed-on: http://review.whamcloud.com/2028
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
James Simmons [Wed, 30 Nov 2011 15:07:18 +0000 (10:07 -0500)]
LU-322 tests: Pool test fixes for large number of OSTs
This patch fixes issues with large numbers of OSTs with the
ost-pools test. We need to use the hexidecimal numbers for
the pool args since the OSTs UUID are named with hex numbers.
Curently the test do pass but errors can be seen in the logs
when more than 9 OSTs exist.
Change-Id: I2bed80a1c55fc7cd405dc530876ef517a380c423
Signed-off-by: James Simmons <uja.ornl@gmail.com>
Reviewed-on: http://review.whamcloud.com/251
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Minh Diep <mdiep@whamcloud.com>
Jinshan Xiong [Thu, 13 Oct 2011 23:27:08 +0000 (16:27 -0700)]
LU-760 obdecho: initialization and awk problem
echo_client registers obd type and then initializes kmem cache, this is
problematic because echo_key will be immidiately accessed after obd type
is registered. This will cause kernel fault.
Also, an awk problem is fixed. If the length of output buffer is longer
than 1024 bytes, awk will run into problem.
Change-Id: Ief287e8f4eeb6a39bc336e7a9f5c21e921b79a58
Signed-off-by: Jinshan Xiong <jay@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1521
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Bobi Jam [Mon, 9 Jan 2012 10:23:41 +0000 (18:23 +0800)]
LU-966 mdd: mdd object may not be exist
If MDT device has been checked with fsck, some mdd objects
could be removed, so that recovery replay could act on non-existing
objects.
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Change-Id: If3a3b72ee70ab2513978ed968c9598ddde11c085
Reviewed-on: http://review.whamcloud.com/1928
Reviewed-by: Fan Yong <yong.fan@whamcloud.com>
Reviewed-by: Mikhail Pershin <tappro@whamcloud.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Oleg Drokin [Mon, 30 Jan 2012 00:27:20 +0000 (19:27 -0500)]
LU-1042 ost: revert negative timestamp initialization
This reverts commit
414251797ed178eec5d431e1f5aa4a889d2b159f from
"LU-221 don't use a/c/m time for newly allocated objects",
because it caused 1.8 clients to think that these files had
a time in the distant past.
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Signed-off-by: Oleg Drokin <green@whamcloud.com>
Niu Yawei [Wed, 11 Jan 2012 03:59:10 +0000 (19:59 -0800)]
LU-919 obdclass: remove hard coded 0x5a5a5a
We assert atomic_t value with hard coded 0x5a5a5a in several places,
which could result in false assertion failure when the reference count
getting very large in some extreme case.
The hard coded 0x5a5a5a should be replaced by LI_POISON.
Signed-off-by: Bruno Faccini <bruno.faccini@bull.net>
Signed-off-by: Niu Yawei <niu@whamcloud.com>
Change-Id: Idc271621017d071b3e2dce5d0ec6fb854127a955
Reviewed-on: http://review.whamcloud.com/1953
Reviewed-by: Lai Siyao <laisiyao@whamcloud.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
James Simmons [Fri, 13 Jan 2012 16:11:29 +0000 (11:11 -0500)]
LU-987 build: Fail to create ldisk rpms
The autoMakefile.am in ldsikfs does not define the
BUILD_SERVER flag so make rpms fails. This patch
simply set the flag to true since ldiskfs will most
likely be used only on servers
Signed-off-by: James Simmons <uja.ornl@gmail.com>
Change-Id: I51e31d1ce0fd1d7e8639426852af1888a9a93f4f
Minh Diep [Thu, 5 Jan 2012 17:12:29 +0000 (09:12 -0800)]
LU-955 build: fix bad lustre-backend-fs dependency
Fix an incorrect RPM package dependency if Lustre RPMs are built
with "make rpms" with client only
Signed-off-by: Minh Diep <mdiep@whamcloud.com>
Change-Id: Ib8691a37fc230cc63b7aca48bc5146a67e10a2f0
Minh Diep [Fri, 13 Jan 2012 23:50:42 +0000 (15:50 -0800)]
LU-843 iokit: get_ec_devno returns trash
Some commands leak output to stdout and get_ec_devno returns
its results via stdout. Redirect fd 8 to stdout and redirect
stdout to stderr. Echo final results to fd 8. Since this
only works when run as a subshell, invoke "exit" instead of
"return" when done.
Inside create_objects() function, stdin is redirected for a
shell "while" command. Though this is done in the current
environment (process) for current BASH, it is not required
and there is no committment that BASH will always behave
this way. Instead, set fd 3 to be reading from that file
and read from fd 3 insted of stdin.
While there, also test for "is object id" lines and extract
values using shell constructs instead of fork-ed commands.
Also while there, ensure that "ERROR" is always printed on
every error exit.
Also while there, ensure that the libecho file comes from
the same directory as the obdfilter-survey script.
Signed-off-by: Minh Diep <mdiep@whamcloud.com>
Change-Id: If6dec690d3aa96f6ddb7fa2994330aae4f89fec0
Reviewed-on: http://review.whamcloud.com/1967
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: wangdi <di.wang@whamcloud.com>
yangsheng [Thu, 5 Jan 2012 10:30:42 +0000 (18:30 +0800)]
LU-781 kernel: kernel update [RHEL6.2 2.6.32-220]
Add support for RHEL6.2. The version is 2.6.32-220.el6.
Change-Id: Icc03a2f5d8b377aa1b1180ae09056989bbc84a9d
Signed-off-by: Yang Sheng <ys@whamcloud.com>
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Signed-off-by: Oleg Drokin <green@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1986
Tested-by: Hudson
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
yangsheng [Mon, 28 Nov 2011 03:09:38 +0000 (11:09 +0800)]
LU-674 kernel update [RHEL5.7 2.6.18-274.12.1.el5].
Test-Paramaters-Start:
serverdistro=el5
clientdistro=el5
testgroup=regression
Test-Paramaters-End:
Change-Id: I146b64a3e5110af868067c4fb5fefde5270ca2fc
Signed-off-by: Yang Sheng <ys@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1744
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-by: Bobi Jam <bobijam@whamcloud.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Bobi Jam [Thu, 27 Oct 2011 01:51:39 +0000 (09:51 +0800)]
LU-508 ldiskfs: fix race in ext4_ext_walk_space()
we should not access on-disk data (e.g. path->p_ext->*) with no
locking.
to be fixed in mainline ext4 as well.
Port from: ORI-291
Author: Alex Zhuravlev <bzzz@whamcloud.com>
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Change-Id: Ic9dfcc9a90766c79c0da5ca54e9fbb2f917865a6
Reviewed-on: http://review.whamcloud.com/1618
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: Mikhail Pershin <tappro@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Jinshan Xiong [Fri, 19 Aug 2011 22:24:15 +0000 (15:24 -0700)]
LU-613 clio: Client dead-lock during binary exec
The root cause is clio takes attr lock and i_size_sem in reverse
order.
In my patch, I find out there is no deadlock issue any more in fault
and truncate path, so it holds i_size_sem in fault path to fix this
problem.
Change-Id: I04cca9324158a34fded6651692410e29aae2e402
Signed-off-by: Jinshan Xiong <jay@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1281
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Niu Yawei <niu@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
hongchao.zhang [Thu, 23 Jun 2011 03:24:31 +0000 (11:24 +0800)]
LU-357 racer test cleanup
1, increase the test time to 300s(900s for SLOW)
2, fixing the problem of recursively calling racer.sh
Change-Id: I91ac7e5c42ed5bc98b3a647c30d7e37af0573f09
Signed-off-by: Hongchao Zhang <hongchao.zhang@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/905
Reviewed-by: Yu Jian <yujian@whamcloud.com>
Reviewed-by: Fan Yong <yong.fan@whamcloud.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Niu Yawei [Wed, 9 Nov 2011 03:23:52 +0000 (19:23 -0800)]
LU-745 kernel: ost-pools test_23 hung
It could be caused by a jbd2 bug which result in forever sleep
in the do_get_write_access().
http://www.spinics.net/lists/linux-ext4/msg24689.html
In do_get_write_access() we wait on BH_Unshadow bit for buffer to get
from shadow state. The waking code in journal_commit_transaction() has
a bug because it does not issue a memory barrier after the buffer is moved
from the shadow state and before wake_up_bit() is called. Thus a waitqueue
check can happen before the buffer is actually moved from the shadow state
and waiting process may never be woken. Fix the problem by issuing proper
barrier.
Signed-off-by: Niu Yawei <niu@whamcloud.com>
Change-Id: I44dce352babc6699cdacc00263bfd3f24538400c
Reviewed-on: http://review.whamcloud.com/1675
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Niu Yawei [Sat, 22 Oct 2011 08:16:55 +0000 (16:16 +0800)]
LU-753 obdfilter: improper LASSERT in filter_commitrw_write()
In rare cases fsfilt_commit_wait() will wake up and return after the
transaction has finished its work and updated j_commit_sequence but
the commit callbacks have not been run yet. Which will trigger the
LASSERT(oti->oti_transno <= obd->obd_last_committed) improperly.
We should just wait for the commit callback finished instead of put
an improper LASSERT here.
Signed-off-by: Niu Yawei <niu@whamcloud.com>
Change-Id: Ibd5add8d352d2e7598be49b0bf8fa37d40ce6e1f
Reviewed-on: http://review.whamcloud.com/1583
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Hudson
Reviewed-by: Jinshan Xiong <jay@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Niu Yawei [Tue, 25 Oct 2011 11:31:11 +0000 (04:31 -0700)]
LU-791 obdfilter: Don't clear OBD_MD_FLFLAGS mistakenly
Instead of set oa->o_valid to OBD_MD_FLID | OBD_MD_FLGROUP arbitrarily
in filter_handle_precreate(), it should be changed as "|=" to keep
the OBD_MD_FLFLAGS setted in filter_precreate().
Otherwise, client will not be aware of OST is running out of space,
and lov_create() will wait for objects forever in such case.
Signed-off-by: Niu Yawei <niu@whamcloud.com>
Change-Id: I84be8dde59dbb2829cd800e10b7aa6f4402b7e56
Reviewed-on: http://review.whamcloud.com/1589
Reviewed-by: Fan Yong <yong.fan@whamcloud.com>
Reviewed-by: hongchao.zhang <hongchao.zhang@whamcloud.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
yangsheng [Thu, 3 Nov 2011 18:25:42 +0000 (02:25 +0800)]
LU-762 ldiskfs: don't drop directory nlink to 0
When landing the nlink patch for ext4, for an unknown reason the
logic in ext4_dec_count() was changed from the ext3 version of
the patch. It now drops the nlink = 0 temporarily and then
if it is a directory with nlink == 0 it increases nlink again.
Instead, only drop nlink if it is larger than 2.
Change-Id: Ieeff3e45daea56f502848f9c2b0fb04f0a9d2b6d
Author: Andreas Dilger <adilger@whamcloud.com>
Signed-off-by: Yang Sheng <ys@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1644
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Lai Siyao [Fri, 8 Jul 2011 16:35:47 +0000 (09:35 -0700)]
LU-488 ptlrpc_connection_put() LASSERT(!cfs_hlist_unhashed(&conn->c_hash))
Connection hash may be rehashed while ptlrpc_connection_put() is
called, ASSERT &conn->c_refcount > 1 instead of this.
Signed-off-by: Lai Siyao <laisiyao@whamcloud.com>
Change-Id: Iec6d35419e0c4d8497bd0b84c6210abc8eb23882
Reviewed-on: http://review.whamcloud.com/1074
Tested-by: Hudson
Reviewed-by: Li Wei <liwei@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Prakash Surya <surya1@llnl.gov>
Reviewed-by: zhen liang <liang.zhen@live.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Christopher J. Morrone [Wed, 20 Jul 2011 00:43:15 +0000 (17:43 -0700)]
LU-513 Make cfs_wait_event_interruptible_exclusive really exclusive
Change-Id: Iea0556a006f8826f8597824131fb5110a848c434
Signed-off-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-on: http://review.whamcloud.com/1118
Tested-by: Hudson
Reviewed-by: Lai Siyao <laisiyao@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Liang Zhen <liang@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Bobi Jam [Fri, 4 Nov 2011 07:22:41 +0000 (15:22 +0800)]
LU-417 llite: report non-zero blocks on writing client
Writing client may not report accurate allocated block numbers when
dirty pages has not been writting back to OSTs, some "cp" or "tar" may
skip the file because it thinks it is completely sparse.
This patch makes writing client consider dirty pages when reporting
allocated blocks, lest the file be treated as a completely sparse one.
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Change-Id: I985d50a44ea1e917bf8e1cba3b5cb770eec35c3f
Reviewed-on: http://review.whamcloud.com/1647
Tested-by: Hudson
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Bobi Jam [Tue, 23 Aug 2011 03:34:06 +0000 (11:34 +0800)]
LU-601 mdd: Fix transaction credits
* mdd_create()/mdd_create_data() may need delete orphan objects on
OSTs, so that we need preserve enough transaction credits for llog
records.
* mdd_attr_set() may write lov llogs.
* orphan_object_destroy() also will write a llog record, we need
reserve credit for it as well.
* add credit changelog record.
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Change-Id: I5124d2f368e2ff794b2b2b8194bec86f63e971cf
Reviewed-on: http://review.whamcloud.com/1276
Reviewed-by: Fan Yong <yong.fan@whamcloud.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Mikhail Pershin <tappro@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Jinshan Xiong [Wed, 5 Oct 2011 05:54:48 +0000 (22:54 -0700)]
LU-687 clio: retry if fault page was truncated
In vvp_io_fault_start, if a page was truncated we should retry it in
ll_fault() instead of return -EFAULT because it will cause fake OOM.
Change-Id: Ia2ba40ca4ecd67170e6c2eb81ddc0ee34d9379a8
Signed-off-by: Jinshan Xiong <jay@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1453
Tested-by: Hudson
Reviewed-by: Niu Yawei <niu@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Fan Yong <yong.fan@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Michael MacDonald [Tue, 25 Oct 2011 13:57:16 +0000 (09:57 -0400)]
LU-792 lbuild-rhel5 should use redhat's SRPM repo
Download EL5 kernel .src.rpm packages from redhat's repo instead
of trying to use CentOS's often-outdated repo.
Change-Id: I0e622e6481f11941b5038b36c0fd074fe470c595
Signed-off-by: Michael MacDonald <mjmac@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1590
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Brian J. Murrell <brian@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Niu Yawei [Wed, 23 Mar 2011 06:30:33 +0000 (23:30 -0700)]
LU-110 Recognize old mdt.quota_type param
Minor changes in mdt_process_config() to make it capable of recognizing old mdt.quota_type.
Issue:
Signed-off-by: Niu Yawei <niu@whamcloud.com>
Change-Id: Ie13851eee573eb633a68bf5ee3cd86dbbf1bd1a9
Reviewed-on: http://review.whamcloud.com/354
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Bobi Jam [Mon, 3 Oct 2011 04:44:37 +0000 (12:44 +0800)]
LU-721 llite: Parallel writes to same file results in zero file
Canceling lock may contain data being sent to OSTs. Change
find_cbdata iterator to take that into account.
Author: Vladimir Saveliev <vladimir.saveliev@oracle.com>
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Change-Id: I1b30fe13ee5303fe3b98e45ad969f5273c41d676
Reviewed-on: http://review.whamcloud.com/1460
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Oleg Drokin [Fri, 16 Sep 2011 17:20:21 +0000 (13:20 -0400)]
Update release date for 2.1 in Changelog to reflect RC2
Signed-off-by: Oleg Drokin <green@whamcloud.com>
Change-Id: Ifda31666cac43cb964184903a3d3de20f671cb7d
Jinshan Xiong [Thu, 15 Sep 2011 05:49:21 +0000 (22:49 -0700)]
LU-683: handle short write in cl_io_commit_write
CLIO has to handle short write case where no bytes were
actually copied from user space. In this case, we should
avoid adding that page into cache.
Change-Id: I78c870e741d7bdb2a9c4c0bfba1adc50a6e814e7
Signed-off-by: Jinshan Xiong <jay@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1383
Tested-by: Hudson
Tested-by: Yu Jian <yujian@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Reviewed-by: Bobi Jam <bobijam@whamcloud.com>
Mikhail Pershin [Wed, 7 Sep 2011 06:26:31 +0000 (10:26 +0400)]
LU-653 i_version shouldn't be used for VBR
RHEL6 plus EXT4 combination cause i_version is to be used to store
Lustre versions instead of i_fs_version. Meanwhile this field is used
by ext4 internally and may be changed breaking Lustre recovery.
Patch makes only i_fs_version to be used for versions.
Change-Id: Idf7cd257aa2d288abce2a7c18c4d8b139bc9c5c6
Signed-off-by: Mikhail Pershin <tappro@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1342
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Tested-by: Yu Jian <yujian@whamcloud.com>
Oleg Drokin [Thu, 1 Sep 2011 20:19:11 +0000 (16:19 -0400)]
LU-653 Ignore last transno from clients with no outstanding transactions
Clients that did not have any unreplied transactions should not
artificially reduce next_recovery_transno since those transactions are
never coming anyway and would just elay all replay activity
until all clients are reconnected.
Change-Id: Iafa32b315211ada93838f4b16fe0f2c800d619c4
Signed-off-by: Oleg Drokin <green@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1318
Tested-by: Hudson
Reviewed-by: Mikhail Pershin <tappro@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Hongchao Zhang [Fri, 9 Sep 2011 11:00:39 +0000 (19:00 +0800)]
LU-662 fix conflict between seq_client_flush and seq_client_alloc_fid
seq_client_flush modifies the lu_client_seq.lcs_space by holding the
lu_client_seq.lcs_sem, but seq_client_alloc_fid uses lcs_update to
exclude other concurrent modification.
Change-Id: Ibc3335841b9a045360283e15aab9b6c36b7a39af
Signed-off-by: Hongchao Zhang <hongchao.zhang@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1364
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
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>
Bobi Jam [Sun, 2 Oct 2011 05:18:59 +0000 (13:18 +0800)]
LU-650 Max possible file size is mutable
Move maxbytes fields out of wire data part of lov_stripe_md,
it can change when OSTs change their status, e.g. disconnected
or inactivated.
Change-Id: I9fb1f2597df1e98d993792706ac94f469d23a78c
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1327
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Hudson
Tested-by: Yu Jian <yujian@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Andreas Dilger [Fri, 26 Aug 2011 08:26:51 +0000 (02:26 -0600)]
LU-614 tests: speed up needlessly long tests
Fix unlinkmany to use the original $next_id instead of the one
seen after createmany was run.
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: I816081ab1ee04fe66d6c14b76eb076262215bc32
Reviewed-on: http://review.whamcloud.com/1298
Reviewed-by: Yu Jian <yujian@whamcloud.com>
Tested-by: Yu Jian <yujian@whamcloud.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Niu Yawei [Fri, 26 Aug 2011 02:58:36 +0000 (19:58 -0700)]
LU-617 temporary fix for the setattr from open
The setattr from open(open(O_TRUNC)) is serialized with
'cl_setattr_lock' on client and goes to a dedicate portal, which is
different with other reint operations, consequently, setattr RPC
can be parallel with other reint RPCs, and that result in the race of
updating last_transno/last_xid on server.
This temporary fix removes the LASSERT(req_is_replay(req)) in
mdt_last_rcvd_update().
Signed-off-by: Niu Yawei <niu@whamcloud.com>
Change-Id: I7e07eaf9e718c2d75d9a503207523be9c700ea4a
Reviewed-on: http://review.whamcloud.com/1295
Tested-by: Hudson
Reviewed-by: Mikhail Pershin <tappro@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Bobi Jam [Sat, 20 Aug 2011 10:51:33 +0000 (18:51 +0800)]
LU-594 Make sure sanity 27v fallout clears
We need wait long enough for OST to wake up from
OBD_FAIL_TGT_DELAY_PRECREATE sleep.
Change-Id: I8e4e8153a11a70b66a64f4e8a1deb3e741f58554
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1262
Reviewed-by: Yu Jian <yujian@whamcloud.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Bobi Jam [Wed, 24 Aug 2011 04:07:22 +0000 (12:07 +0800)]
LU-622 Alloc enough cfg buffer space
obd device name needs space enough for "<dev_name>-<instance_name>\0".
This fixes the glitch in commit
cbc4ca2e8dc37d54bb7d3c9b02ab20b63e60f592.
Change-Id: Ib8c3c8db49baad8a9ca068a1c8009aee72aa1e62
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1280
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Reviewed-by: Jinshan Xiong <jay@whamcloud.com>
Reviewed-by: Yu Jian <yujian@whamcloud.com>
Tested-by: Yu Jian <yujian@whamcloud.com>
Andreas Dilger [Fri, 19 Aug 2011 21:58:43 +0000 (15:58 -0600)]
LU-614 test: speed up needlessly long tests
Some tests are running needlessly long compared to the value that
they provide, and can be made much faster without any loss in value.
conf-sanity test_53a: run llverdev in partial mode always
conf-sanity test_53b: run llverfs in partial mode always
sanity test_220: use unlinkmany instead of "rm -rf", but could likely
be sped up dramatically via CFS_FAIL_LOC() on OST
Change-Id: Ifbe66bffc051b6163e41a9bffb1d6b1ab6304e94
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1266
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Yu Jian <yujian@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Oleg Drokin [Tue, 23 Aug 2011 03:54:16 +0000 (23:54 -0400)]
Update release date for 2.1.0 RC1
Change-Id: Ia71c1f4a4d6867e0e605675a1d57a541ef5f9c92
Signed-off-by: Oleg Drokin <green@whamcloud.com>
Oleg Drokin [Sat, 20 Aug 2011 16:52:43 +0000 (12:52 -0400)]
LU-365 Update lustre/Changelog
To indicate that the release is from Whamcloud.
Change-Id: I2e8f8e83c895ef1591d55b460727e04a695e1f29
Signed-off-by: Oleg Drokin <green@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1269
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Oleg Drokin [Sat, 20 Aug 2011 16:30:26 +0000 (12:30 -0400)]
Update version to 2.1 as a start of RC process.
Signed-off-by: Oleg Drokin <green@whamcloud.com>
Change-Id: I46e6405953581b5cc10140e8f25e2b4b345e1a4d
Oleg Drokin [Sat, 20 Aug 2011 16:29:52 +0000 (12:29 -0400)]
Update recommended e2fsprogs version
Signed-off-by: Oleg Drokin <green@whamcloud.com>
Change-Id: I00a587a4ab4e8fecb6d3def33ee952b2f793873e
Oleg Drokin [Fri, 19 Aug 2011 23:56:06 +0000 (19:56 -0400)]
LU-365 Update copyright for files modified by Whamcloud
I scoured through the commits logs since first whamcloud commit
and updated all the files that had non-trivial contributions by
Whamcloud.
Signed-off-by: Oleg Drokin <green@whamcloud.com>
Change-Id: I293a82b69a4ce02c028517771bd172ed94c4e44b
Reviewed-on: http://review.whamcloud.com/1267
Reviewed-by: Robert Read <rread@whamcloud.com>
Tested-by: Hudson
nasf [Tue, 16 Aug 2011 13:53:19 +0000 (21:53 +0800)]
LU-598 open non-exist object should return ENOENT
When call ll_intent_file_open() against non-exist object, we cannot
obtain RPC request refcount after the call, so subsequent operation
should not use such internal (only) RPC request to avoid to access
freed (or re-allocated) memory.
Change-Id: I4c3763f460c6145e376538f392fde338d44e6c5e
Signed-off-by: nasf <yong.fan@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1242
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andreas Dilger [Wed, 29 Jun 2011 20:36:25 +0000 (14:36 -0600)]
LU-473 llite: improve error for 2.x cli + 1.8 srv
Improve the error message printed to the console when a 2.x client
is trying to mount a 1.8 filesystem. More specifically, have the 2.x
client check whether the MDS supports the OBD_CONNECT_FID and
OBD_CONNECT_FULL20 flags, without which the client will not be able
to cope. The 2.x client also does not understand the RPC structure
of 1.8 servers (mdt_body, mdt_rec, etc), but this is harder to check.
For a 2.x server, the client will configure an LMV device, so
exp_connect_flags will be zero, but in that case we also know the
server is running 2.x.
Change-Id: I935571aa0623f9311ddf0545904230721d2fad32
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1036
Tested-by: Hudson
Reviewed-by: Yu Jian <yujian@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Niu Yawei [Tue, 16 Aug 2011 09:38:20 +0000 (02:38 -0700)]
LU-592 1.8 <-> 2.1 interop: missing LMV on client
When the cluster is upgraded from 1.8, client will not have LMV layer,
and the "ll_md_exp" is pointing to MDC directly, so the code which
assuming "ll_md_exp" as LMV need be revised:
- ll_dir_ioctl(LL_IOC_GETOBDCOUNT) should return 1 for MDC case;
- mdc_iocontrol() needs to handle the OBD_IOC_QUOTACTL as well when the
"ll_md_exp" is MDC;
Signed-off-by: Niu Yawei <niu@whamcloud.com>
Change-Id: I3659a5f8076125c9b2012826521c44540e65a75e
Reviewed-on: http://review.whamcloud.com/1236
Tested-by: Hudson
Reviewed-by: Fan Yong <yong.fan@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
nasf [Wed, 17 Aug 2011 07:18:34 +0000 (15:18 +0800)]
LU-599 verify "mi_ginfo" before accessing its member
If the target user only belong to one group, then we just set
"identity->mi_gid", and does not set "identity->mi_ginfo", to
save unnecessary memory allocation and groups processing.
So before accessing member inside "identity->mi_ginfo", make
sure "identity->mi_ginfo" is valid.
Change-Id: I880ccadddd93425c301115328d57c552d11bda8e
Signed-off-by: nasf <yong.fan@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1252
Tested-by: Hudson
Reviewed-by: Yu Jian <yujian@whamcloud.com>
Reviewed-by: hongchao.zhang <hongchao.zhang@whamcloud.com>
Tested-by: hongchao.zhang <hongchao.zhang@whamcloud.com>
Tested-by: Yu Jian <yujian@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Bobi Jam [Sat, 13 Aug 2011 14:53:22 +0000 (22:53 +0800)]
LU-585 lower mdt reconstruct open debug message level
Reconstruct open is not an error case.
Change-Id: Ifb559616c148b01298196b4be7e398306a6b8e4c
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1233
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
hongchao.zhang [Fri, 12 Aug 2011 13:25:13 +0000 (21:25 +0800)]
LU-325 using preallocated objects if OST has enough disk space
during creating files, if ENOSPC is returned by one OST, the preallocated
objects in the corresponding OSC should still can be used if there is enough
blocks but no inode available in the OST.
Change-Id: I9d57461698d1c0fb3a492188fb4ce1519d65efc7
Signed-off-by: Hongchao Zhang <hongchao.zhang@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1051
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Fan Yong <yong.fan@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Lai Siyao [Thu, 21 Jul 2011 21:04:42 +0000 (05:04 +0800)]
LU-586 1.8 <-> 2.1 interop: list_add corruption
In seq_client_alloc_fid(), client may list_add itself more than
once to waitqueue under race or error condition.
Signed-off-by: Lai Siyao <laisiyao@whamcloud.com>
Change-Id: Id7d35cc4ca3eafc4fc44d2d79a9ebc25fe56e62c
Reviewed-on: http://review.whamcloud.com/1216
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-by: Liang Zhen <liang@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
yangsheng [Wed, 20 Apr 2011 10:05:51 +0000 (18:05 +0800)]
LU-15 Strange slow IO messages and bad performance
b=24183 slow I/O on new files via mballoc.
upstream patch to avoid loading bitmaps from full groups
Change-Id: I9b4de1b4b1942b0f084b6199d5ab3e1267c9e8e3
Signed-off-by: Yang Sheng <ys@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/442
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Yu Jian [Thu, 11 Aug 2011 10:34:27 +0000 (18:34 +0800)]
LU-544 speed up filesystem formatting time in conf-sanity.sh
Pass "-E lazy_itable_init" to mke2fs to speed up Lustre
filesystem formatting time in conf-sanity.sh.
Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: I29070e6c4b95028b12573d552a32ca04189c2c1d
Reviewed-on: http://review.whamcloud.com/1210
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Christopher J. Morrone [Fri, 22 Jul 2011 20:28:06 +0000 (13:28 -0700)]
LU-455 Replace DIST_SOURCES with EXTRA_DIST
Resolve autoconf-2.63 warnings mainly by replacing DIST_SOURCES
with EXTRA_DIST. Additionally, the non-portable $(shell) contruct
was removed from tests/module.mk and the sources simply enumated.
Finally, the incorrect EXTRA_SOURCES instance was removed from
the top level autoMakefile.am.
Running automake-1.11 -a -c -W no-portability
libcfs/libcfs/autoMakefile.am:92:
variable `DIST_SOURCES' is defined but no program or
libcfs/libcfs/autoMakefile.am:92:
library has `DIST' as canonical name (possible typo)
lnet/klnds/mxlnd/autoMakefile.am:44:
variable `DIST_SOURCES' is defined but no program or
lnet/klnds/mxlnd/autoMakefile.am:44:
library has `DIST' as canonical name (possible typo)
...
Also, as a result of the above changes, libcfs/libcfs/autoMakefile.am had to be
modified in order to allow 'make dist' to succeed.
As it turns out, libcfsutils_a_SOURCES was incorrect. That list contained
references to nonexistent 'util/[parser|platform].h' files. The assumption is
it intended to reference libcfs/include/libcfs/util/[parser|platform].h.
To fix the issue, both [parser|platform].h references were removed from the
list. This produces a simple solution that maintains consistency with the rest
of the build system. The caveat being that libcfsutil.a won't automatically be
rebuilt if either of the intended [parser|platform].h files are modified.
Change-Id: Ia81eb1e3fc219f6dac4c7da234f7e736754c5440
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Christopher J. Morrone <morrone2@llnl.gov>
Signed-off-by: Prakash Surya <surya1@llnl.gov>
Reviewed-on: http://review.whamcloud.com/1092
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Richard Henwood <rhenwood@whamcloud.com>
Tested-by: Richard Henwood <rhenwood@whamcloud.com>
Reviewed-by: Brian J. Murrell <brian@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Niu Yawei [Fri, 29 Jul 2011 04:00:09 +0000 (21:00 -0700)]
LU-550 incorrect quota hash function
Quota code uses 'uid' or '(1 << 32) + gid' as the hash key, however,
the quota hash function lqs_hash() always try to interpret such a
unsigned long long key as a 'quota_adjust_qunit' struct, then calculate
the hash value by quota_adjust_qunit->qaq_id.
Signed-off-by: Niu Yawei <niu@whamcloud.com>
Change-Id: I48dc7f276eb97f34b22097b4edc529d5f76acf37
Reviewed-on: http://review.whamcloud.com/1162
Tested-by: Hudson
Reviewed-by: Fan Yong <yong.fan@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Yu Jian [Thu, 14 Jul 2011 03:07:34 +0000 (11:07 +0800)]
LU-297 improve llverfs to display the current read/write performance
This patch changes show_filename() to show_rate() in llverfs.c to
display the current read/write file name and performance, along with
an estimate of how long the whole read/write operation will continue.
The patch also changes "fast" to "partial" and fixes some typos in
the comments of llverdev.c and llverfs.c.
Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: I94cb112ae9a38b1128e53021aa0c1f1e8e5411af
Reviewed-on: http://review.whamcloud.com/1110
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
John Hammond [Mon, 1 Aug 2011 15:28:48 +0000 (10:28 -0500)]
LU-560 Fixup intent status on ENOENT in mdt_intent_reint
Clear -ENOENT from rep->lock_policy_res2 for non existent files so
that clients may cache negative dentries if they hold
MDS_INODELOCK_UPDATE on the parent directory.
Change-Id: I238f1001782923a595f9f7ee1c094d53ab1422ae
Signed-off-by: John Hammond <jhammond@tacc.utexas.edu>
Reviewed-on: http://review.whamcloud.com/1173
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andreas Dilger [Tue, 9 Aug 2011 21:17:02 +0000 (15:17 -0600)]
LU-574 Fix whitespace issues in conf-sanity.sh
Move "run_one test_41b" to immediately follow declaration of test_41b.
Fix some whitespace problems (space before tab, space at end of line)
that I noticed while inspecting LU-574.
Change-Id: I18a80a55c9895cf8a6647a01ffa646f5ca499795
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1200
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
James Simmons [Fri, 13 May 2011 13:25:01 +0000 (09:25 -0400)]
LU-321 Don't assume file system name for some test
Some of the test assume the file system name is lustre
which is not always the case.
Change-Id: I582aa81fce5d1a881837dfa28bb0d6d84cfb5547
Signed-off-by: James Simmons <uja.ornl@gmail.com>
Reviewed-on: http://review.whamcloud.com/253
Tested-by: Hudson
Reviewed-by: Brian J. Murrell <brian@whamcloud.com>
Reviewed-by: Yu Jian <yujian@whamcloud.com>
Tested-by: Yu Jian <yujian@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Jinshan Xiong [Sat, 30 Jul 2011 07:24:51 +0000 (00:24 -0700)]
LU-556: Fix config_log_find() in mgc_request.c
This issue was imported by lu-411. Actually after checking
the code, I found we can make things a bit simpler.
Change-Id: Iad212782a5f0357b73014baf7766c77637fb47db
Signed-off-by: Jinshan Xiong <jay@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1168
Tested-by: Hudson
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Mikhail Pershin <tappro@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Yang Sheng <ys@whamcloud.com>
Liang Zhen [Mon, 1 Aug 2011 13:07:58 +0000 (21:07 +0800)]
LU-292 srv_n_queued_reqs is screwed up by ptlrpc_unregister_service
service::srv_n_queued_reqs is counter for pre-unpack request on
service::srv_req_in_queue, we shouldn't change it for unpacked request
returned by ptlrpc_server_request_get()
Change-Id: I770411fa8f986fa387a1790f87e6a265abc469f1
Signed-off-by: Liang Zhen <liang@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1172
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andreas Dilger [Fri, 29 Jul 2011 10:59:12 +0000 (04:59 -0600)]
LU-553 add checkpatch.pl for Lustre coding style
Add a modified version of the Linux kernel checkpatch.pl script (from 2.6.38)
to Lustre for validating patches against the Lustre coding style. The Lustre
coding style is largely the same as the kernel, except using spaces instead of
tabs for indenting.
Change-Id: I2d116c6769f9211825dc8ddaead05914f2116c60
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1164
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Niu Yawei [Thu, 4 Aug 2011 06:31:10 +0000 (23:31 -0700)]
LU-559 Keep the o_seq unchanged for 1.8 client
ost_validate_obdo() arbitrarily change the o_seq to 0 for the requests
from 1.8 client, which will make the 1.8 echo client use the 0 group
mistakenly. In fact, I think we needn't do anything special for the 1.8
client here.
Signed-off-by: Niu Yawei <niu@whamcloud.com>
Change-Id: I93cd3bdbf984f8aef0f5b68f5ee8640f31ded337
Reviewed-on: http://review.whamcloud.com/1182
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: wangdi <di.wang@whamcloud.com>
nasf [Tue, 2 Aug 2011 16:26:18 +0000 (00:26 +0800)]
LU-540 misc patch for user identity upcall/downcall
1) l_getidentity forgets to allocate supplementary group info space, but access
such non-allocated memoey without checking.
2) In structure "identity_downcall_data", "idd_gid" is same as "idd_groups[0]",
remove such redundant information.
3) More sanity check for lprocfs_wr_identity_info().
4) Drop unnecessary memory allocation to simplify mdt_identity_do_upcall().
5) Add padding field in "identity_downcall_data" to resolve align issues.
6) Use new magic to distinguish old version/format downcall.
Change-Id: Ib0dfb37c39a7cb335e29557b68b885afa726789e
Signed-off-by: nasf <yong.fan@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1160
Tested-by: Hudson
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andreas Dilger [Wed, 27 Jul 2011 08:30:19 +0000 (02:30 -0600)]
LU-540 quota: add compatibility for 1.8 userspace
The LL_IOC_QUOTACTL ioctl data structure if_quotactl changed size
in 2.x by adding new fields in the middle of the structure. Add
compatibility for the old 1.8 quotactl userspace ioctl so that it
is possible for existing quotactl tools to run on both 1.x and 2.x.
This requires renumbering the LL_IOC_QUOTACTL ioctl, since it had
incorrectly used "struct if_quotactl *" as the "size" parameter,
but the _pointer_ size did not change. ioctl() values should always
take a struct as the size parameter, not a pointer to the struct,
so that issues like this can be caught earlier.
Change-Id: I0bce98327a276c0c9a49ce78efd3f45e00df7168
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1152
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>
Christopher J. Morrone [Mon, 8 Aug 2011 12:30:44 +0000 (08:30 -0400)]
LU-533 Include posix-types.h in headers and distribution
Include libcfs/posix/posix-types.h in the main rpm, and
add #include statements in lustre_user.h and ll_fiemap.h.
This will allow user-space applications that include these
files to compile.
Change-Id: Ibe391bf28afe91b3914ca1eb07fe66e408b21b4a
Signed-off-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-on: http://review.whamcloud.com/1138
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
yangsheng [Fri, 5 Aug 2011 09:54:40 +0000 (17:54 +0800)]
LU-429 Fix and quiet debug message in filter_connect
The comment refers to mds_connect(), which no longer exits. So I
just removed the comment.
The message says "Received MDS connection", but this function is
used when clients connect to OSTs as well. We don't need to see
this message on the console all of the time, so I changed it to
D_INFO.
Change-Id: Iebfbaafe39df702862ade126b979528855173d5c
Signed-off-by: Christopher J. Morrone <morrone2@llnl.gov>
Signed-off-by: Yang Sheng <ys@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/963
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
hongchao.zhang [Thu, 28 Jul 2011 05:43:59 +0000 (13:43 +0800)]
LU-350 fix bug in mgs to allow starting OST before MDT
in mgs_steal_llog_for_mdt_from_client, it forgets to add the "index" and
"svname" of the existing OST to the config log of the MDT's lov, then cause
Lustre fails to mount if the OSTs start before MDT.
Change-Id: I0e68bf51c88517dbd42790e7e4447a186cf268f3
Signed-off-by: Hongchao Zhang <hongchao.zhang@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/595
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Fan Yong <yong.fan@whamcloud.com>
Reviewed-by: Yu Jian <yujian@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
John Hammond [Thu, 14 Jul 2011 00:48:40 +0000 (19:48 -0500)]
LU-333: Make read_bytes in llite/*/stats report bytes read.
Move ll_stats_ops_tally(..., LPROC_LL_{READ,WRITE}_BYTES, ...) from
vvp_io_init() to the end of ll_file_io_generic(), and tally by the
actual number of bytes read rather than the number requested.
Test 127b, client stats are correctly tallied with this patch.
Change-Id: Icad0bc09630b8c1775f2ec59326902aa88215498
Signed-off-by: John Hammond <jhammond@tacc.utexas.edu>
Signed-off-by: Richard Henwood <rhenwood@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1094
Tested-by: Hudson
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Richard Henwood [Wed, 6 Jul 2011 20:21:01 +0000 (16:21 -0400)]
LU-487 brw_[read|write] stats to correctly counts number of pages.
brw_[read|write] now correctly counts the number of pages. Before
this change, the number of pages multiplied by the page size was
counted.
Change-Id: I069c18cf46c67285de1ae2afc48f7fb4ab92df3c
Signed-off-by: Richard Henwood <rhenwood@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1056
Tested-by: Hudson
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Niu Yawei [Thu, 28 Jul 2011 10:47:46 +0000 (03:47 -0700)]
LU-530 group quotas not enforced
In filter_quota_getflag(), we shouldn't break the loop when the file's owner
doesn't have quota limit, otherwise, the check for file's group will be skipped
and the out of quota flag will not be returned to client at the end.
Signed-off-by: Niu Yawei <niu@whamcloud.com>
Change-Id: I3e92ba4c061409d5a911d46916535c71f7843e0b
Reviewed-on: http://review.whamcloud.com/1156
Tested-by: Hudson
Reviewed-by: Fan Yong <yong.fan@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Bobi Jam [Thu, 28 Jul 2011 10:11:54 +0000 (18:11 +0800)]
LU-541 Fix liblustreapi.c usage of errno and rc.
* fix liblustreapi.c usage of errno and rc, llapi_xxx() functions should
return negative value on error.
* enable sanity subtest 65.
Change-Id: I30d172c49b60f6f77ea57c3082d0c68f4c141dea
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1155
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: hongchao.zhang <hongchao.zhang@whamcloud.com>
Tested-by: hongchao.zhang <hongchao.zhang@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
nasf [Sun, 31 Jul 2011 16:26:46 +0000 (00:26 +0800)]
LU-540 remove file flags definition conflict between 1.8 and 2.x
There is file flags definition conflict between 1.8 "LL_FILE_LOCKED_DIRECTIO"
and 2.x "LL_FILE_RMTACL". Redefine the later one to erase such conflict.
Change-Id: I1f4b96382f865f8548bf1a54fb1c74d3649b0daf
Signed-off-by: nasf <yong.fan@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1159
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Oleg Drokin [Tue, 2 Aug 2011 03:08:19 +0000 (23:08 -0400)]
LU-451 Workaround for lockd (nfsv3) locking
It seems that lockd works in a different way lock-wise compared
to everybody else. They actually don't use lock owner to determine
conflicts and install their own owner compare that compares specially
crafted pids (duh!).
As a workaround if a custom owner compare is detected, assign pid to
owner field (not intersecting with kernel pointer space anyway).
svclock is the only place using custom owner compare function ATM.
Change-Id: Ia89e2327705f2afbe200fdbadd0f86038fda936d
Signed-off-by: Oleg Drokin <green@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1174
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Lai Siyao [Wed, 15 Jun 2011 02:52:55 +0000 (19:52 -0700)]
LU-290 Reconnects are not throttled
bz22423
Client should not try to reconnect to the same nid in a busy loop,
but instead rely on pinger to issue reconnect.
Signed-off-by: Lai Siyao <laisiyao@whamcloud.com>
Change-Id: I878093030d62173071192ca816e037464cc9248d
Reviewed-on: http://review.whamcloud.com/944
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-by: hongchao.zhang <hongchao.zhang@whamcloud.com>
Tested-by: hongchao.zhang <hongchao.zhang@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Christopher J. Morrone [Thu, 9 Jun 2011 20:14:11 +0000 (13:14 -0700)]
LU-407 Fix lustre-modules rpm name in spec file.
RHEL and Fedora kernels include ".%{_arch}" in the kernel's version
string. Since Lustre includes the kernel version in its rpms, it
can appear that a double architecture name appears at the end. Really
one is from the kernel, and one is from Lustre. Trying to change that
is probably more trouble than it is worth.
This seems to cause problems with rpm's understanding of rpm names. To
fix that this patch takes Andreas's suggestion of grep'ing the full
rpm name from "rpm -q kernel-modules".
We also add "|| true" to the last line of the %preun scriptlet so that
failure does not aport installation.
Finally, we also remove the redundant greps in the postun scriptlet.
Change-Id: I2c71e853e28ec6e0907eb4ea7c3205ca6e5dd873
Signed-off-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-on: http://review.whamcloud.com/934
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Brian J. Murrell <brian@whamcloud.com>
hongchao.zhang [Thu, 28 Jul 2011 07:46:47 +0000 (15:46 +0800)]
LU-327 cleanup the client import of mgc
the client import will not have been cleaned if it is setup but never
connected, adding such cleanup code in mgc_precleanup.
define a new inline function "obd_cleanup_client_import" in "obd_class.h"
to be used by mgc, mdc and osc.
Change-Id: If033de3fe81b9a565a7af40a7b93e395fb0bd659
Signed-off-by: Hongchao Zhang <hongchao.zhang@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/917
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Fan Yong <yong.fan@whamcloud.com>
Reviewed-by: Mikhail Pershin <tappro@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
yangsheng [Wed, 27 Jul 2011 08:01:10 +0000 (16:01 +0800)]
LU-409 Oops: RIP: _spin_lock_irq+0x15/0x40
Skip sanity test_68a since llite_loop module disabled in rhel6.
Change-Id: I213b435e64ac3816f1a3d564522a8c8797fc387c
Signed-off-by: Yang Sheng <ys@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1150
Reviewed-by: Yu Jian <yujian@whamcloud.com>
Tested-by: Hudson
Reviewed-by: Richard Henwood <rhenwood@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Bobi Jam [Tue, 26 Jul 2011 17:36:18 +0000 (01:36 +0800)]
LU-539 obd_data_connect server side could be corrupted
In target_handle_connect(), we use assignment to copy client's connect
data to its server counterpart, and we define the pointers pointing to
a big buffer (struct obd_connect_data), while server's reply connect
data could be a smaller buffer (struct obd_connect_data_v1), the
assignment could corrupt the request's obd_connect_data reply buffer.
Change-Id: Ib72776d3dee8a2a836fda5b29ab49c3bbc026cab
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1144
Reviewed-by: Fan Yong <yong.fan@whamcloud.com>
Tested-by: Hudson
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
hongchao.zhang [Fri, 29 Jul 2011 03:33:13 +0000 (11:33 +0800)]
LU-207 adding stats for some operations in MDT
Adding stats for getattr, setattr, statfs and sync in MDT
Change-Id: I4790bb2eb6bc756a24942224506c7306429243bc
Signed-off-by: Hongchao Zhang <hongchao.zhang@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/427
Reviewed-by: Yu Jian <yujian@whamcloud.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
yangsheng [Fri, 22 Jul 2011 18:50:33 +0000 (02:50 +0800)]
LU-527 Update RHEL6 kernel to 2.6.32-131.6.1 security update
Change-Id: I5a84a5240ed8e4d119ae3094beb575666c69f7c6
Signed-off-by: Yang Sheng <ys@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1135
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andreas Dilger [Wed, 27 Jul 2011 05:25:18 +0000 (23:25 -0600)]
LU-540 lloop: fix LLOOP_INFO to return FID
Instead of returning an inode number to userspace for the LLOOP_INFO
ioctl, instead return a FID value.
This does not pose compatibility problems, because the ioctl() number
was already different between Lustre 1.x and 2.x, and the lctl needed
for using lloop devices is always built with the kernel modules.
Change-Id: Ia3ee6dab830ad65e2a11d9248982c82ce76221f1
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1149
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Jinshan Xiong <jay@whamcloud.com>
Reviewed-by: Fan Yong <yong.fan@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andreas Dilger [Wed, 27 Jul 2011 07:54:20 +0000 (01:54 -0600)]
LU-540 quota: move quota ioctl handling to helper
The ll_dir_ioctl()::OBD_IOC_QUOTACTL handling is very large, and
is moved to a helper routine quotactl_ioctl() instead of being inline.
This will also facilitate the compatibility handling for the 1.x
QUOTACTL ioctl handling, by being able to share quotactl_ioctl() code.
Also, check for LL_IOC_QUOTACTL in ll_dir_ioctl(), since this is the
ioctl() number exported to userspace, and it also facilitates being
able to find the kernel-side ioctl handler for llapi_quotactl().
Add comments to show the duplicate user/kernel ioctl definitions.
Keep these separate for now, to allow the kernel/userspace values
to be defined differently for compatibility reasons.
There should be no functional differences with this change.
Change-Id: Iecfaf20ddbd78d77a546e1ba5a60b281732a85d9
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1151
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-by: Fan Yong <yong.fan@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andreas Dilger [Wed, 13 Jul 2011 03:41:39 +0000 (21:41 -0600)]
LU-20 remove unnecessary DMU export patch
The patch for kernel symbol export for DMU back-ends was added to
facilitate upgrading servers. However, the upstream ZFS code will
no longer be working on RHEL5 kernels, so this patch is not needed.
Change-Id: Icf443a366fd51c06cb5d02df27b7e70f476fcb63
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1086
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Tested-by: Hudson
Reviewed-by: Yang Sheng <ys@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Andreas Dilger [Fri, 17 Jun 2011 20:01:44 +0000 (14:01 -0600)]
LU-228 kernel: remove more unused kernel patches
The filterdata patches have been unused and unnecessary for some
time already. The jbd-slab-race and include-fixes patches were
merged into the upstream kernel and back-ported into RHEL5.
Change-Id: I01851a60ecf453f6eb5dbfc4f573444f4f220861
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/962
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Yang Sheng <ys@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andreas Dilger [Fri, 17 Jun 2011 18:36:23 +0000 (12:36 -0600)]
LU-73 kernel: remove wrong jbd2 commit timer patch
The bug that this patch was originally fixing (incorrectly rounding
down the jiffies timer, which might cause it to be skipped) was
fixed in the upstream kernel by using round_jiffies_up(), which
always moves the timer into the future.
Remove this patch entirely, since it is no longer needed.
Change-Id: I3b4f30e0565cf64b95fc87babf1f54a9ab190754
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/961
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Yang Sheng <ys@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
yangsheng [Fri, 22 Jul 2011 18:47:43 +0000 (02:47 +0800)]
LU-528 Update RHEL5 kernel to 2.6.18-238.19.1 security update
Change-Id: I13974999954ae1806c48f75fa6d35f4e98cf2a75
Signed-off-by: Yang Sheng <ys@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1134
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andreas Dilger [Thu, 14 Jul 2011 19:07:29 +0000 (13:07 -0600)]
LU-278 build: Only warn for tag/version mismatch
The configure process should NOT abort just because the most
recent tag is not of the form that upstream uses to tag Lustre.
Downstream developers may use their own tags, or just add
extensions to upsteam's version tags.
Change-Id: I9a98bfd4475d3df2694f536ba0352779a62650c7
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/510
Tested-by: Hudson
Reviewed-by: Brian J. Murrell <brian@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
yangsheng [Tue, 14 Jun 2011 20:12:06 +0000 (04:12 +0800)]
LU-411 Kernel panic when running conf-sanity on RHEL5/i686
Reduce stack usage to avoid stack overflow on
RHEL5/i686 patchless client.
Signed-off-by: Yang Sheng <ys@whamcloud.com>
Change-Id: I54d3a19734ab089f44f43943b80a2e7ad1f3335f
Reviewed-on: http://review.whamcloud.com/950
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Yu Jian [Wed, 8 Jun 2011 13:23:32 +0000 (21:23 +0800)]
LU-372 add $TMP/mgsactive to indicate the active combined MGS/MDS node
The current test-framework could not figure out the active MGS node
correctly while the MGS and MDS nodes are combined under a failover
configuration. We need add a $TMP/mgsactive file to indicate the
active MGS node like what we did for the MDS node.
Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: I74bbf9293f9a290573c4b9fcbf9fea0fcc0aad95
Reviewed-on: http://review.whamcloud.com/913
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>
hongchao.zhang [Sat, 16 Jul 2011 22:16:55 +0000 (06:16 +0800)]
LU-221 don't use a/c/m time for newly allocated object in OST
in OST, the a/c/m time is unusable for newly allocated object for
it has not gotten valid time info from clients. by setting the a/c/m
time as LONG_MIN at creation will make the clients ignore the time.
Change-Id: I0d6094e4257626ffd1128b3540837deaf230e056
Signed-off-by: Hongchao Zhang <hongchao.zhang@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1084
Tested-by: Hudson
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Niu Yawei <niu@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Jinshan Xiong [Tue, 26 Jul 2011 23:42:34 +0000 (16:42 -0700)]
LU-394: LND failure casued by discontiguous KIOV
Fix the problem of assigning starting_offset twice. This is imported
in combining patchset of lu-394 and lu-523.
Change-Id: Ia4ea205f211c70965101507906b02799a65694be
Signed-off-by: Jinshan Xiong <jay@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1146
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Tested-by: Oleg Drokin <green@whamcloud.com>
Alexey Lyashkov [Wed, 13 Jul 2011 03:16:12 +0000 (07:16 +0400)]
LU-502 don't allow to kill service threads by OOM killer.
OOM can produce a many issues like requests processing counter.
Change-Id: I8d417d148b589ea6de84c5d16bc2607a59852564
Signed-off-by: Alexey Lyashkov <alexey_lyashkov@xyratex.com>
Reviewed-on: http://review.whamcloud.com/1088
Tested-by: Hudson
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Oleg Drokin [Tue, 26 Jul 2011 15:50:15 +0000 (11:50 -0400)]
New version 2.0.66
Change-Id: Ic2aa5252ec569efefd00255440c5465cae633e1c
Signed-off-by: Oleg Drokin <green@whamcloud.com>
Jinshan Xiong [Tue, 12 Jul 2011 17:55:19 +0000 (10:55 -0700)]
LU-394: LND failure casued by discontiguous KIOV
This issue was imported by bug 18881 where I moved the urgent
pages to front of lop_pending to fix a deadlock issue.
I reverted bug 18881 in this patch and came up with a new solution:
cl_page_gang_lookup() only blocks on the first page. This is also
for deadlock avoid since we should never grab multiple pages' lock
without try method.
Change-Id: I5dce35e3929e4f79a350e56ddc9e752269db060e
Signed-off-by: Jinshan Xiong <jay@whamcloud.com>
Signed-off-by: Oleg Drokin <green@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/911
Jinshan Xiong [Sat, 23 Jul 2011 03:52:46 +0000 (20:52 -0700)]
LU-523: No prepare_write for lockless IO
For page unaligned write, CLIO does prepare_write even for lockless IO
and then write full page back, this will cause data corruption since
data is not covered by lock.
In this patch, we don't do prepare_write for lockless IO, and then submit
exact bytes in the page to OST.
Change-Id: I4aa5afeb82cb717de499c8a8c004078b279302c7
Signed-off-by: Jinshan Xiong <jay@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1130
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Reviewed-by: Niu Yawei <niu@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Niu Yawei [Mon, 27 Jun 2011 10:55:10 +0000 (03:55 -0700)]
LU-456 Force commit to reuse the just-deleted blocks
In ext4 non-writeback journal mode, the just-deleted blocks are not
useable until the transaction committed, this could cause filter
return -ENOSPC mistakenly when the just-deleted blocks are not committed.
Whenever block allocation fails for -ENOSPC on filter, we should wait
for the previous deleted blocks committed, then retry the allocation.
Signed-off-by: Niu Yawei <niu@whamcloud.com>
Change-Id: I678b05ae9aa1b9e9d0aa29f48279bc7daedebbb2
Reviewed-on: http://review.whamcloud.com/1022
Tested-by: Hudson
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Johann Lombardi [Tue, 19 Jul 2011 12:15:58 +0000 (14:15 +0200)]
LU-507 check_write_rcs() return -EPROTO instead of -ENOSPC
remote_rcs[i] is an unsigned value so the check remote_rcs[i] < 0 is
always false.
Change-Id: I5468d9185e4c8de819f58693375e9d96722b9fbb
Signed-off-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1115
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Li Wei [Tue, 5 Jul 2011 08:44:36 +0000 (16:44 +0800)]
LU-476 Run run-llog.sh on MGS in sanity 60a
sanity 60a had been effectively skipped during lab testing, in which the
MGS is not running on the host where sanity.sh is executed. This patch
fixes the subtest by requesting run-llog.sh to be run on the MGS
explicitly.
Change-Id: I9dc47d856cd42bfe0cd5b5b71e3fd7d4f7c0e561
Signed-off-by: Li Wei <liwei@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/1050
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Bobi Jam [Tue, 22 Mar 2011 04:41:17 +0000 (12:41 +0800)]
LU-16 Allow objects larger than 2TB in size
Use OST superblock sb->s_maxbytes telling LOV the maxmimum object size
it supports, and LOV records the minimum of the per-OSC maxbytes values
to determine the lov_stripe_md.lsm_maxbytes.
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Change-Id: Ice920dc59f5cc5855393bb7031fa089757b3314d
Bugzilla: 20128
Reviewed-on: http://review.whamcloud.com/241
Tested-by: Hudson
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Yu Jian [Thu, 14 Jul 2011 06:32:14 +0000 (14:32 +0800)]
LU-477 allocate memory for s_group_desc and s_group_info by vmalloc()
Add the patch to the RHEL6 ldiskfs patch series.
Large kmalloc() for sbi->s_group_desc and sbi->s_group_info can fail
for large filesystems, which will cause the "not enough memory" error
while mounting. This patch makes it fall back to vmalloc() if the
kmalloc() failed, as what was done for sbi->s_flex_groups.
To avoid colliding with an valid on-disk inode number, EXT4_BAD_INO
is used as the number of the buddy cache inode.
The patch also incorporates the following upstream kernel fix:
commit
32a9bb57d7c1fd04ae0f72b8f671501f000a0e9f
ext4: fix missing iput of root inode for some mount error paths
https://bugzilla.kernel.org/show_bug.cgi?id=26752
Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: Ia263c90759e96710702e4afff3ba19e77455386f
Reviewed-on: http://review.whamcloud.com/1095
Tested-by: Hudson
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Chris [Thu, 10 Mar 2011 11:05:37 +0000 (11:05 +0000)]
LU-121 Fixup yaml.sh and test-framework.sh so that it only ever uses the first
part of the host name for log files.
This means stripping of everything off the name after and including the first .
This change is only designed to make the .yml files consistent. The log files
will append the fully qualified name and appear to do this consitantly.
We can create a jira to make the log and yml files consistant, but the reality
is that the yml files have a short life and are only used to send to Maloo.
This change will allow automated posting of results and automated testing to
begin.
1. Carries out the above using hostname -s [Thanks Andreas]
2. Adds in a quick change so that LUSTRE_BUILD in yaml.sh can be a reference to
the source rather than just the lustre version string which is recorded and
written to the yaml anyway as LUSTER_VERSION
Additionally a couple of other changes sneaked in.
1. Allows the review information to be applied to the yaml output file
by way of exporting the variable CODE_REVIEW_YAML to be a yaml description
for maloo
2. The addition of a couple of fixes to make the permissions for yaml
files be allow-all. This permissions are bracketed and so do not change any
other parts of the code.
Change-Id: Ied9a2117cd1b68e048bbc2bf1bffe782485c6150
Signed-off-by: Chris Gearing <chris@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/312
Tested-by: Hudson
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Yu Jian <yujian@whamcloud.com>