LU-4222 mdt: extra checking for getattr RPC.

It should check whether getattr RPC can hold
layout MD(RMF_MDT_MD), in case the client sends
some invalid RPC, which can cause panic on MDT.

Client will retrieve cl_max_md_size/cl_default_md_size
from MDS during mount process, so it will initialize
cl_max_md_size/cl_default_md_size before sending getattr
to MDS.

Signed-off-by: wang di <di.wang@intel.com>
Change-Id: Ic12f281d574e7138ee0e1d5aae6161b7607837cc

New tag 2.4.3-RC1

Change-Id: I35b61517da77ed9ca9109fd0be323a84aaf626f6

LU-4101 mdt: protect internal xattrs

In mdt_reint_setxattr() require CAP_SYS_ADMIN to modify a trusted
xattr and silently disallow modification those trusted xattrs used by
Lustre internally.

Signed-off-by: John L. Hammond <john.hammond@intel.com>
Change-Id: Ic616dca74a90da0aedb0ec2624618f91ac6fcaf4
Reviewed-on: http://review.whamcloud.com/7943
Reviewed-by: Andrew Perepechko <andrew_perepechko@xyratex.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>

LU-4704 acl: fix permission problem of setfacl

Setxattr does not check the permission when setting ACL xattrs. This
will cause security problem because any user can walk around
permission checking by changing ACL rules.

Lustre-commit: 42f504ecada81d7a2a8e2244f345e8dbf73fd157
Lustre-change: http://review.whamcloud.com/9473

Signed-off-by: Li Xi <lixi@ddn.com>
Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Emoly Liu <emoly.liu@intel.com>
Change-Id: I5844c5b654afd70fb3b5ee91c024736b377124b0
Reviewed-on: http://review.whamcloud.com/9559
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>

LU-4703 mdd: do not skip xattr sanity check for all cases

xattr sanity check should be done at all times.
Otherwise we are risking letting a non root user setting
access acl on any file.

Lustre-commit: 6d7066989c06bdf6c86f5b8b61d542fda8e5739e
Lustre-change: http://review.whamcloud.com/9469

Signed-off-by: Li Dongyang <dongyang.li@anu.edu.au>
Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Change-Id: Ic763424b2a20dab00c7e62de1f5c8553cd0cdaa8
Reviewed-on: http://review.whamcloud.com/9558
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>

LU-3680 ptlrpc: Fix assertion failure of null_alloc_rs()

lustre_get_emerg_rs() set the size of the reply buffer to zero
by mistake, which will cause LBUG in null_alloc_rs() when memory
pressure is high. This patch fix this problem and adds a size
check to avoid the problem of insufficient buffer size.

Signed-off-by: Li Xi <lixi@ddn.com>
Signed-off-by: Patrick Farrell <paf@cray.com>
Change-Id: I9fbd4f14e8e1263de2af564c4f2e420f5f2b43bc
Reviewed-on: http://review.whamcloud.com/8200
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-4360 Fix use after free in ksocknal_send

Call to ksocknal_launch_packet might schedule a callback that
might free the just sent message, and so subsequent access to it
via lntmsg->msg_vmflush goes to freed memory.

Instead we'll just remember if we are in the vmflush thread and
only restore if we happened to set mempressure flag.

Change-Id: I2f0f8b27e26e11b37ad60fde4c98e86c39768349
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>

New tag 2.4.2-rc2

Change-Id: Ie8a11ab58c0e77b7cb5d4c479ceb6fdaa7987c2d

Revert "LU-3433 clio: wrong cl_lock usage"

This seems to cause 3889

This reverts commit 1a04ba1a772659cc35f806210551ebf542c6365e.

New tag 2.4.2rc1

Change-Id: Ic27617445c7a7b996ba554e9699cdc4d670b8f07
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>

LU-4015 ldiskfs: access external journal using device path

Currently ldiskfs when configured with an external journal
uses the device number to access it. The problem with this
approach is that the device number can change across
reboots. This patch adds a new mount option -o journal_path
to tell ldiskfs where the external journal is instead. This
patch is based on commit landed to what will be Lustre 2.6
which is d406aa3e71cde588ea53d42ff2596d77e65e33f8.

Signed-off-by: James Simmons <uja.ornl@gmail.com>
Change-Id: I86af198cdab54164583ae1c03fff891edfc1f152
Reviewed-on: http://review.whamcloud.com/8471
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Nunez <james.a.nunez@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-3433 clio: wrong cl_lock usage

granted lock is moved to HELD state in cl_wait_try() only which is
done after upcall for non-agl locks. as the result, lock unuse moves
cl_lock not to CACHED state, but to NEW state. A parallel thread
gets this lock and tries to enqueue it - instead of re-using a
cached lock, cl_enqueue_try() initiates a new enqueue and gets to
osc_lock_enqueue() where it asserts:

LASSERTF(ols->ols_state == OLS_NEW, ...);

the state of osc lock is RELEASED already - moved here by unuse.

This patch is back-ported from the following one:
Lustre-commit: 521335cefe670efa2dc34c5db522a283f318447e
Lustre-change: http://review.whamcloud.com/6709

Signed-off-by: Vitaly Fertman <vitaly_fertman@xyratex.com>
Change-Id: I27904bb6198efd5ff98e01f2f74767beabf03a45
Signed-off-by: Jian Yu <jian.yu@intel.com>
Reviewed-on: http://review.whamcloud.com/8530
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-4365 quota: wait for global lock cancel

In qsd_qtype_fini(), we'd wait for the global lock cancel done.

Test-Parameters: envdefinitions=SLOW=yes,ENABLE_QUOTA=yes \
mdtfilesystemtype=zfs mdsfilesystemtype=zfs ostfilesystemtype=zfs \
testlist=recovery-small

Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Change-Id: Id235c7d6e07f5ce436655a6d5382e4c8c161fa3b
Reviewed-on: http://review.whamcloud.com/8586
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2800 llite: introduce local getname()

Commit 8e377d15 (v3.7) unexported getname and putname. Although
getname was re-exported, putname was not. All we really need is
something to to allocate a buffer and copy it from userspace so
let's just use our own version. The down size is that we lose
the audit support in the original getname.

Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Signed-off-by: yang sheng <yang.sheng@intel.com>
Change-Id: If44cd9f9d14f26f220d5f0c52e8b804a49d3cf92
Reviewed-on: http://review.whamcloud.com/8581
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-4362 libcfs: wrong parameter to cfs_hash_keycpy

cfs_hash_rehash_key() passed wrong parameter to cfs_hash_keycpy,
hnode should be the second parameter not the third one.

Signed-off-by: Liang Zhen <liang.zhen@intel.com>
Change-Id: Ib5b61a2f2bbd85848429ae5eae3702cb98d548d8
Reviewed-on: http://review.whamcloud.com/8510
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Johann Lombardi <johann.lombardi@intel.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>

LU-3929 lwp: don't connect LWP to old MDT

When rolling upgrade Lustre from old version without LWP support
(< 2.3.60) to newer version, the OSTs will be upgraded firstly,
then LWP will try to connect the old MDT, and old MDT will regard
the LWP client as normal client and save client export on disk for
recovery, which will break recovery at the end since LWP
connection isn't recoverable.

Change-Id: Iad6de8f8a59ff9606c30dd8b750e7638b6bc4e9c
Signed-off-by: Hongchao Zhang <hongchao.zhang@intel.com>
Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Reviewed-on: http://review.whamcloud.com/8513
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-4313 osp: continue waiting for -ENOTCONN

in osp_precreate_reserve, -ENOTCONN should also be treated
as recoverable errors, such as -ENODEV, and continue waiting
the object to be created by OSP once it completes recovery.

This patch is back-ported from the following one:
Lustre-commit: 3a2073d3bf9fab33fa26c5c5f0da2b0821185131
Lustre-change: http://review.whamcloud.com/8415

Signed-off-by: Hongchao Zhang <hongchao.zhang@intel.com>
Change-Id: I1fd13c8bec9b7827396ec922cb985ee642a5e286
Signed-off-by: Jian Yu <jian.yu@intel.com>
Reviewed-on: http://review.whamcloud.com/8468
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-4243 mgs: use proper uuid in mgs_write_log_failnids()

This patch revert the changes to mgs_write_log_failnids() in the
commit d9d27cad, which introduced a regression of that when multiple
'--servicenode' or '--failnode' option specified, the uuid from
first option will be wrongly used for all other connections.

Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Change-Id: Ibd989de67fe6851cc75569be3f64dadaec5e6f8c
Reviewed-on: http://review.whamcloud.com/8372
Tested-by: Jenkins
Reviewed-by: Erich Focht <efocht@gmail.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-3906 tests: use df instead of lfs_df in run_compilebench()

Since run_compilebench() will be run on NFS client which has no
Lustre filesystem, we need change to use df instead of lfs_df to
get the space usage information.

Test-Parameters: envdefinitions=SLOW=yes,ONLY=compilebench \
testlist=parallel-scale,parallel-scale-nfsv3,parallel-scale-nfsv4

Signed-off-by: Jian Yu <jian.yu@intel.com>
Change-Id: Ib3bf1eebf4175b9a80308d1b5d4ff9c7926eb068
Reviewed-on: http://review.whamcloud.com/8430
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Reviewed-by: Wei Liu <wei3.liu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-3316 obdclass: fix wrong device put in case of race

The ls_device_put() might be called from local_oid_storage_fini()
wrongly if mutex-protected check failed due to race

This patch is back-ported from the following one:
Lustre-commit: 458804418f3c24706db7ad9a46f791e59f008a7d
Lustre-change: http://review.whamcloud.com/6334

Signed-off-by: Mikhail Pershin <mike.pershin@intel.com>
Change-Id: I386597e431ded0dd2b63b078200571a00bc5fe96
Signed-off-by: Jian Yu <jian.yu@intel.com>
Reviewed-on: http://review.whamcloud.com/8461
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-4175 tests: fix obdecho_test() to calculate write count

This patch fixes obdecho_test() in sanity.sh to calculate
write count so as to avoid out of space failure while the
page number is large but the obd target size is small.

Signed-off-by: Jian Yu <jian.yu@intel.com>
Change-Id: Ifcd055c587570d3f0e3ca22d763f9478e9bee3fa
Reviewed-on: http://review.whamcloud.com/8388
Tested-by: Jenkins
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Emoly Liu <emoly.liu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-3582 test: Add run_test to Runtests

There is no run_test in this script so add one
so logging in Autotest will greatly improve.
It seems some of the test is dependant so I did a
very minimal change to get the better logging.

The patch is back-ported from the following patch:
Lustre-commit: c74be05495cd5b7489fdbb100cd7b6d15db32053
Lustre-change: http://review.whamcloud.com/7014

Signed-off-by: Keith Mannthey <keith.mannthey@intel.com>
Change-Id: I96d7fe464d81094842dc5939519d74f7168d96b1
Signed-off-by: Jian Yu <jian.yu@intel.com>
Reviewed-on: http://review.whamcloud.com/8384
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Emoly Liu <emoly.liu@intel.com>
Reviewed-by: Wei Liu <wei3.liu@intel.com>

LU-4310 kernel: missed unlock in quota patch

The quota-replace-dqptr-sem.patch introduced a regression which
could miss unlock inode->i_lock in __dquot_alloc_space().

This patch is back-ported from the following patch:
Lustre-commit: 1fa094b5977820127ccae4ea7c13904c9716c6fb
Lustre-change: http://review.whamcloud.com/8400

Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Change-Id: I2353a8f60ea0b4d54332eaafcb2397f6c0ed28e6
Signed-off-by: Jian Yu <jian.yu@intel.com>
Reviewed-on: http://review.whamcloud.com/8442
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2829 tests: decrease load on zfs for sanityn/33a-b

This patch decreases the number of files again for sanityn
test 33a and 33b based on the following patch:

Lustre-commit: 41ba272d1904bb9c612f785f608fc2d396763581
Lustre-change: http://review.whamcloud.com/8273

Test-Parameters: envdefinitions=SLOW=yes \
mdtfilesystemtype=zfs mdsfilesystemtype=zfs ostfilesystemtype=zfs \
ostcount=2 testlist=sanityn

Signed-off-by: Jian Yu <jian.yu@intel.com>
Change-Id: Ia768d13b8db174e46ecc284433989a17cd560d07
Reviewed-on: http://review.whamcloud.com/8439
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-4219 mdd: limit os_namelen to the max of NAME_MAX

While running pathconf(".", _PC_NAME_MAX) on ldiskfs-based
Lustre filesystem, it returns 255 (NAME_MAX). However, on
zfs-based Lustre filesystem, it returns 256 (MAXNAMELEN).
Since Lustre can not handle the name length large than
NAME_MAX, this patch limits the value of os_namelen in
mdd_statfs() to the max of NAME_MAX.

Test-Parameters: envdefinitions=SLOW=yes \
mdtfilesystemtype=zfs mdsfilesystemtype=zfs ostfilesystemtype=zfs \
ostcount=2 testlist=posix

Lustre-commit: 7d197837e0d7eb5f40a928a8f6c57270493b8996
Lustre-change: http://review.whamcloud.com/8217

Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Signed-off-by: Jian Yu <jian.yu@intel.com>
Change-Id: I6de2f4c45f33c362b5ff61575f0ebc75477c6b48
Reviewed-on: http://review.whamcloud.com/8276
Tested-by: Jenkins
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>

LU-3738 tests: improve posix.sh to support BASELINE_FS=zfs

This patch improves posix.sh to support BASELINE_FS=zfs. It also
adds create_zpool(), create_zfs() and destroy_zpool() common
functions into test-framework.sh.

Test-Parameters: envdefinitions=SLOW=yes \
mdtfilesystemtype=zfs mdsfilesystemtype=zfs ostfilesystemtype=zfs \
ostcount=2 testlist=posix

Signed-off-by: Jian Yu <jian.yu@intel.com>
Change-Id: Ib8b3cf988bd7066a84a6f9f50ccc159ffeab6e0e
Reviewed-on: http://review.whamcloud.com/8323
Tested-by: Jenkins
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Nathaniel Clark <nathaniel.l.clark@intel.com>

LU-1233 tests: clean up files in performance-sanity test

This patch cleans up the files created in performance-sanity.sh
through mdsrate-{create,lookup,stat}-*.sh after create/lookup/stat
operation fails.

Test-Parameters: envdefinitions=SLOW=yes,ENABLE_QUOTA=yes \
testlist=performance-sanity

Signed-off-by: Jian Yu <jian.yu@intel.com>
Change-Id: I6d51b847fcc9c514c4eab72cd8d422527bdbbc71
Reviewed-on: http://review.whamcloud.com/8289
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Emoly Liu <emoly.liu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-14 tests: add version check code to conf-sanity test 69

This patch adds Lustre version check codes into conf-sanity
test 69 to make the test work with servers which do not have
the following patch:

Lustre-commit: db6613f5bed1606cc8f97b46d1b298746af03a75
Lustre-change: http://review.whamcloud.com/7443

Test-Parameters: envdefinitions=SLOW=yes,ENABLE_QUOTA=yes,ONLY=69 \
ossjob=lustre-b2_3 mdsjob=lustre-b2_3 ossbuildno=41 mdsbuildno=41 \
testlist=conf-sanity

Signed-off-by: Jian Yu <jian.yu@intel.com>
Change-Id: I43e12fda5c0c83848b96bd7f377b7b4514ed5d5e
Reviewed-on: http://review.whamcloud.com/8412
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>

LU-3935 scrub: support dryrun mode OI scrub

The administrator can detect whether there are any inconsistent
OI mappings but without repairing them via the command:

lctl lfsck_start --dryrun on -M <$device>

Under default mode, dryrun is "off". When the device is mounted,
the initial OI scrub will detect whether there are upgrade case,
or file-level backup/resotre case, or OI file(s) recreated case.
If found some inconsistency, it will trigger non-dryrun mode OI
scrub automatically. If the administrator does NOT want to such
auto non-dryrun OI scrub to avoid some unexpected repairing, he
(or she) can specify "-o noscrub" when mount the device.

Test-Parameters: testlist=sanity-scrub
Signed-off-by: Fan Yong <fan.yong@intel.com>
Change-Id: Ibab79c543dbb5b69789b688cacca10737a710f83
Reviewed-on: http://review.whamcloud.com/7867
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-4179 mdt: skip open lock enqueue during resent

Skip open lock enqueue, if the open lock has been
acquired(mdt_intent_fixup_resent) during resent.

Signed-off-by: wang di <di.wang@intel.com>
Change-Id: I625ca438e28520416ee2af884d0a9f9e6f21cf2e
Reviewed-on: http://review.whamcloud.com/8173
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-3879 test: Check for e2fsprogs on server

sanity-quota checks if e2fsprogs is loaded on the client,
but should be checking for e2fsprogs on the server. This
patch changes checking for dumpe2fs on the client to
checking on the MDS.

Signed-off-by: James Nunez <james.a.nunez@intel.com>
Change-Id: I85d3b71dedb6e33775c7823b11f18c7ca405e858
Reviewed-on: http://review.whamcloud.com/8350
Tested-by: Jenkins
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-4016 tests: use user and group names in metadata-updates

pdsh eats %u passed to stat executed on remote nodes.  Avoid this
problem by using the user and group names instead of numerical
identifiers in metadata-updates ownership checks.

Test-Parameters: testlist=metadata-updates
Signed-off-by: James Nunez <james.a.nunez@intel.com>
Signed-off-by: Stephen Champion <schamp@sgi.com>
Change-Id: Id0a283d58873bb64772f28cbc28c5a6abad9a48e
Reviewed-on: http://review.whamcloud.com/8052
Tested-by: Jenkins
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Reviewed-on: http://review.whamcloud.com/8333

LU-14 utils: allow formatting OST without VIRGIN flag

When reformatting a new target to replace an existing target, the
"mkfs.lustre --replace" option allows formatting an OST (or MDT?)
with the same index as a previously used OST without setting the
LDD_F_VIRGIN flag.

This is required in the case where an OST was lost due to massive
corruption or critical hardware failure.  Otherwise, the newly
formatted target will try to register with the MGS as a new target,
but will be refused by the MGS due to having an already-used index.

The OFD code skips precreating all of the objects in the filesystem
if the MDS requests an object ID too much larger than the current
LAST_ID.  In this case (which can happen if the OST is replaced
or restored from an older backup) only the most recent objects are
precreated, and the deletion of orphans from a too-large precreate
is left to lfsck.

Lustre-commit: db6613f5bed1606cc8f97b46d1b298746af03a75
Lustre-change: http://review.whamcloud.com/7443

Test-Parameters: testlist=conf-sanity envdefinitions=SLOW=yes

Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Sebastien Buisson <sebastien.buisson@bull.net>
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Change-Id: I5b773bd7116df4d7875ff31bdd0c873377f83116
Reviewed-on: http://review.whamcloud.com/8159
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>

LU-992 doc: remove old kernels from lustre/ChangeLog

Remove the old and obsolete kernels from the current list of
client and server kernel versions that Lustre works with.

Update recommended e2fsprogs version to include "or newer".

Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Change-Id: I164b7649a9957f9b6882eec45d275651663ebbe5
Reviewed-on: http://review.whamcloud.com/8189
Tested-by: Jenkins
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-3906 tests: fix space estimation codes in run_compilebench()

While running compilebench test, the actual space required for each
kernel directory is about 1GB instead of 680MB. This patch fixes this
number in run_compilebench() so as to make the space estimation
more correct.

Test-Parameters: envdefinitions=SLOW=yes,ONLY=compilebench \
testlist=parallel-scale,parallel-scale-nfsv3,parallel-scale-nfsv4

Test-Parameters: envdefinitions=SLOW=yes,ONLY=compilebench \
mdtfilesystemtype=zfs mdsfilesystemtype=zfs ostfilesystemtype=zfs \
ostcount=2 \
testlist=parallel-scale,parallel-scale-nfsv3,parallel-scale-nfsv4

Signed-off-by: Jian Yu <jian.yu@intel.com>
Change-Id: Ic19e2bff352d3f7444f1da30cedfc832b825873f
Reviewed-on: http://review.whamcloud.com/8288
Tested-by: Jenkins
Reviewed-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Reviewed-by: Wei Liu <wei3.liu@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2887 tests: disable SLOW for ZFS in parallel-scale

Because all zfs pools are on the same physical device in autotest
ZFS backed OSTs run VERY slowly so disable SLOW. Decrease load on zfs
backed MDS.

This patch also allows overriding all variables that are set lower
due to SLOW=no, which is forced for ZFS tests.  This will allow values
specified by caller for all variables even when testing on ZFS.

The patch is back-ported from the following two patches:
Lustre-commit: 474bbd262b095cee441e24ee0818745678236d9c
Lustre-change: http://review.whamcloud.com/7778
and
Lustre-commit: abd7b8c9e13681fd3baf463ccb9619f9bdf25e30
Lustre-change: http://review.whamcloud.com/8234

Test-Parameters: envdefinitions=SLOW=yes,ENABLE_QUOTA=yes \
mdtfilesystemtype=zfs mdsfilesystemtype=zfs ostfilesystemtype=zfs \
ostcount=2 testlist=parallel-scale

Signed-off-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Signed-off-by: Jian Yu <jian.yu@intel.com>
Change-Id: Ifd05ae4ef876bbaf3d7fe0caf38288cd973e716c
Reviewed-on: http://review.whamcloud.com/8284
Tested-by: Jenkins
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>

LU-2829 tests: Decrease load on zfs for sanityn/33a-b

Decrease number of files when testing sanityn/33a and 33b when using
ZFS due to performance issues (cf LU-2887).  ZFS performance when
commit_on_sharing is enabled is very slow.

Lustre-commit: 4add39c3775a69c2b3daf212b8e3fc9976bd53e6
Lustre-change: http://review.whamcloud.com/8161

Test-Parameters: ostcount=2 envdefinitions=SLOW=yes testlist=sanityn mdsfilesystemtype=zfs mdtfilesystemtype=zfs ostfilesystemtype=zfs

Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Signed-off-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Change-Id: I79b6b6c315b7f349120301db514aaf84b7dfb94f
Reviewed-on: http://review.whamcloud.com/8273
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>

LU-4223 tests: delete loop device when umount

Delete loop device with umount -d, to avoid loop device leaking.

Signed-off-by: wang di <di.wang@intel.com>
Change-Id: I55bafefa1d387116333aafc71b599ef888c57a2b
Reviewed-on: http://review.whamcloud.com/8296
Tested-by: Jenkins
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>

LU-3240 mdc: Check for all attributes validity in revalidate

GETATTR needs to return attributes protected by different bits, so
we need to ensure all we have locks with all of those bits, not
just UPDATE bit.

Xyratex-bug-id: MRP-1052
Signed-off-by: Alexey Lyashkov <alexey_lyashkov@xyratex.com>
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Change-Id: I0e27a33341c424a8666bdbf3cdb83b1b4fe0058f
Reviewed-on: http://review.whamcloud.com/6460
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: wangdi <di.wang@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>

LU-3701 ldlm: refine LU-2665 patch for POSIX compliance

Follow-on to patch introduced to fix LU-2665 ticket (Gerrit
Change at http://review.whamcloud.com/6415 with Change-Id:
I8faa331712abeadee46eabe111ee1c23a05840d5).
Original patch introduced regressions against POSIX test suite
(fcntl.18/fcntl.35 tests in LSB-VSX POSIX test suite at
http://www.opengroup.org/testing/linux-test/lsb-vsx.html),
so the idea is to only resend F_UNLCKs to have both LU-2665 bug
and POSIX test suite happy.

Test-Parameters: testlist=posix
Signed-off-by: Bruno Faccini <bruno.faccini@intel.com>
Change-Id: I8d325cb7039bb3f3f74f2a9829cbaea760d9e20f
Reviewed-on: http://review.whamcloud.com/7453
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2665 mdc: Keep resend FLocks

FLocks requests (particulary F_UNLCKs) can't be trashed
upon comm problems with Server/MDS nor upon kill/exit,
thus we need to keep retry/send.

Signed-off-by: Bruno Faccini <bruno.faccini@intel.com>
Change-Id: I8faa331712abeadee46eabe111ee1c23a05840d5
Reviewed-on: http://review.whamcloud.com/6415
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-4139 quota: improve write performance when over softlimit

To improve the write performance when over softlimit, we choose
to lose some grace time accuracy: shrinking the qunit size to
(4 * least_qunit_size) instead of least_qunit_size.

Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Change-Id: I20e341d53979ac42227197ec6fa04dd38974073f
Reviewed-on: http://review.whamcloud.com/8078
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Johann Lombardi <johann.lombardi@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>

LU-3166 build: Fix failed OFED event checking on OFED-3.5

compat driver's header for OFED-3.5 is missing to check
whether OFED events are defined in OFED-3.5.

backport from http://review.whamcloud.com/#/c/6048/

Signed-off-by: Shuichi Ihara <sihara@ddn.com>
Change-Id: If7f888e4343e94021270dac79028ded46226ea6e
Reviewed-on: http://review.whamcloud.com/8205
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-3812 tests: check Lustre version in sanity-scrub test 10b

This patch adds Lustre version check codes into sanity-scrub
test 10b to make the test work with servers which have the
following patch:

Lustre-commit: f5491211ed7251a45e4e6916539cd2a2ab5e6038
Lustre-change: http://review.whamcloud.com/6343

Test-Parameters: envdefinitions=SLOW=yes,ENABLE_QUOTA=yes \
clientdistro=el6 serverdistro=el6 clientarch=x86_64 \
serverarch=x86_64 serverjob=lustre-master serverbuildno=1666 \
testlist=sanity-scrub

Signed-off-by: Jian Yu <jian.yu@intel.com>
Change-Id: I66646a27fc30399fcb718a961362800988d7175e
Reviewed-on: http://review.whamcloud.com/7529
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Wei Liu <wei3.liu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-3763 utils: set multipath devices recursively

Make set_blockdev_tunables() set tunables recursively to reach
slaves' slaves.

Lustre-commit: aeb887ab0d4b15d17f3b96bf0e7be3b746edcf8e
Lustre-change: http://review.whamcloud.com/7337

Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Change-Id: I8f8e8fb1e3500591e33fc745ea9d6cdc05043be8
Reviewed-on: http://review.whamcloud.com/8163
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>

LU-3587 tests: add sanity test for crash in changelog

This patch adds a sanity test to verify that the previous fixes committed
for this bug prevent a crash during a changelog fetch after renaming
one very long filename to another.

Test-Parameters: testlist=sanity envdefinitions=ONLY=160b

Lustre-commit: 3f2ad881802092ef359aa3cf2040978829d444c4
Lustre-change: http://review.whamcloud.com/7751

Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Change-Id: Ia2c9147dcb88f00c56df3e72bb0738c8142c034f
Reviewed-on: http://review.whamcloud.com/8171
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>

LU-3587 changelogs: Correct KUC code max changelog msg size

The kernel to userspace communication routines (KUC) allocate
and limit the maximum cs_buf size to CR_MAXSIZE.  However this
fails to account for the fact that the buffer is assumed to begin
with a struct kuc_hdr.  To allocate and account for that space,
we introduce a new define, KUC_CHANGELOG_MSG_MAXSIZE.

Lustre-commit: b1bc221e1eb98314513eff07afc0b4477076c959
Lustre-change: http://review.whamcloud.com/7406

Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Signed-off-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: jacques-Charles Lafoucriere <jacques-charles.lafoucriere@cea.fr>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Change-Id: I67068dc6e2d73df15dce7fe00957e4e204aa7e59
Reviewed-on: http://review.whamcloud.com/8170
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>

LU-2374 mkfs: Make order of --index and --ost irrelevant

The order of --index and --ost should not prevent mkfs.lustre to
work. So reordered the checks which ensure that  mkfs with index is
done for ost and mdt only.

Lustre-commit: 3052583f5e89b4b4294a4df05939e749be22df3d
Lustre-change: http://review.whamcloud.com/5942

Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Signed-off-by: Girish Shilamkar <gshilamkar@ddn.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Change-Id: I042bbd0ab2157b397c21095857e2011a18f54c98
Reviewed-on: http://review.whamcloud.com/8164
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Nunez <james.a.nunez@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-4118 kernel: kernel update [RHEL6.4 2.6.32-358.23.2.el6]

Update RHEL6.4 kernel to 2.6.32-358.23.2.el6.

Lustre-commit: eee860761cbc2f78bacc040cf36a20137d3ddfe7
Lustre-change: http://review.whamcloud.com/8053

Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Signed-off-by: yang sheng <yang.sheng@intel.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Change-Id: I3ca049d77ad3fe0b1f441fbf45681cfb0c3e0519
Reviewed-on: http://review.whamcloud.com/8166
Reviewed-by: James Simmons <uja.ornl@gmail.com>

LU-3497 build: Use alt. path for ZFS development headers

In the move to dkms the ability to use ZFS userland developement
headers from the ZFS source tree instead of /usr/include was
removed. This patch restores the ability to build lustre using
ZFS on any build box that lacks having the ZFS developement rpms
installed.

Signed-off-by: James Simmons <uja.ornl@gmail.com>
Change-Id: I901b00cfa27798e571d6579e81f191f0a79536da
Reviewed-on: http://review.whamcloud.com/8261
Tested-by: Jenkins
Reviewed-by: Minh Diep <minh.diep@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-3460 quota: don't schedule adjust when qsd stopped

When the qsd is stopped, we should not schedule quota adjust or
index update anymore, because the writeback thread has been stopped,
and those scheduled requests will never be processed.

Lustre-commit: 76ddf87488010a38a3e7b0b5923e0fe8e725326a
Lustre-change: http://review.whamcloud.com/6731

Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Reviewed-by: Johann Lombardi <johann.lombardi@intel.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Change-Id: I7c10c79db4176b3af8c7741dd438a2095ddb9bb5
Reviewed-on: http://review.whamcloud.com/8169
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>

LU-3430 nrs: Fix a race condition in the ORR policy

Checking the atomic oo_ref in nrs_orr_hop_put_free() and then
taking the bucket lock leaves a race window open, by which a
second thread can attempt to free an already-freed
nrs_orr_object. Fix this, and change the hash bucket lock type
from an rwlock to a spinlock, as there are now as many calls on
the write path, as there are on the read path. Rehashing is not
used on the hashes, so libcfs_hash API usage can also be
simplified.

Lusrtr-commit: 548b8ea916625b3697b0e0d0abfc330b4c954633
Lustre-change: http://review.whamcloud.com/7623

Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Signed-off-by: Nikitas Angelinas <nikitas_angelinas@xyratex.com>
Xyratex-bug-id: MRP-1294
Reviewed-by: Liang Zhen <liang.zhen@intel.com>
Reviewed-by: Emoly Liu <emoly.liu@intel.com>
Reviewed-by: Li Xi <pkuelelixi@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Change-Id: I08365b9888d781cb2bcbf871bf74d089bcf7ba9f
Reviewed-on: http://review.whamcloud.com/8162
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>

LU-3829 ldlm: MDT mount fails on MDS w/o MGS on it

If we specify multiple --mgsnode for a MDT, when we start MDS upon
it while MGS is no the other node, the MGC import connection will
always select the local nid (which is one of the candidate mgsnode)
since it think its the closest connection.

This patch treats further --mgsnode nids as failover nids, so that
multiple import connections are added for the MGC import.

Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Change-Id: Ifd015a6df47d4285ff5081a9e1e6eabfc405b193
Reviewed-on: http://review.whamcloud.com/7509
Reviewed-by: Liang Zhen <liang.zhen@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-4007 libcfs: fix for removing lcw at lcw_dispatch_main

lcw_dispatch_main() creates zombies list and moves lcw into it.
But removes entry from another list in the zombies loop.

Signed-off-by: Alexander Boyko <alexander_boyko@xyratex.com>
Xyratex-bug-id: MRP-1179
Change-Id: I9daec3d9b3dbbf244a28b829cd724c8ec60c4b6c
Reviewed-on: http://review.whamcloud.com/7755
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Tested-by: Jenkins
Reviewed-by: Doug Oucharek <doug.s.oucharek@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Liang Zhen <liang.zhen@intel.com>
Reviewed-by: Isaac Huang <he.huang@intel.com>

LU-3173 build: Incorrect detection of ldiskfs series to use

Extract correct upstream kernel release for RHEL kernels and
form it into RHEL_KERNEL_VERSION as
$(VERSION).$(PATCHLEVEL).$(SUBLEVEL)-$(RHEL_RELEASE)
Then use this string for correct ldiskfs series selection.

Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Change-Id: Idda0a4b02a5090b7be6fb81b8f98175559e575aa
Reviewed-on: http://review.whamcloud.com/6056
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>

LU-3587 changelogs: Account for changelog_ext_rec in CR_MAXSIZE

CR_MAXSIZE needs to account for an llog_changelog_rec that actually
contains a changelog_ext_rec structure rather than a changelog_rec.
With out doing so, a file size approaching the Linux kernel NAME_MAX
length that is renamed to a size also close to, or at, NAME_MAX will
exceed CR_MAXSIZE and trip an assertion.

Change-Id: I2d86955efcbd6d0fd28c87026a7720311b4f5598
Signed-off-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-on: http://review.whamcloud.com/6993
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2842 mdt: Do not lbug on invalid fid entry in name

Instead of crashing MDS when the name exists but has invalid fid,
we should just return error and let lfsck deal with it later.

Change-Id: Iea19dac75e9dc75c43a18efa757fc1c5e4d1743e
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Reviewed-on: http://review.whamcloud.com/7617
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: wangdi <di.wang@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>

LU-3161 lod: Added proper error handling instead of LASSERT

LASSERT are replaced with proper error handling in
lod_initialize_objects.

Signed-off-by: Gaurav Mahajan <gmahajan@ddn.com>
Change-Id: Ia8fbbb4956765835bf46b46b3a7addc99cb191fa
Reviewed-on: http://review.whamcloud.com/7234
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>

LU-3903 osd-zfs: Remove dead property handling code

The udmu_userprop_get_str(), udmu_userprop_set_str(), and
udmu_objset_name_get() functions are all currently unused.
Therefore this code is removed from the zfs osd.

If manipulation of zfs dataset properties is needed in the
future it can be done through the existing dsl_prop_get(),
dsl_prop_set_int(), and dsl_prop_set_string() interfaces.

This patch allows Lustre to cleanly build with versions
of zfs newer than 0.6.2.

Change-Id: I15b1ef8825dcdbed1572990d230776485d627457
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-on: http://review.whamcloud.com/7685
Reviewed-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Tested-by: Hudson
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>

LU-3027 lov: to not modify lov lock when sublock is canceled

Otherwise it will cause wrong lock to be enqueued.

lls_ever_canceled is introduced to solve potential performance
problem by this patch.

Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Change-Id: I1ea62902c9d1a468f5952f30a8dc5dec22b8bb1c
Reviewed-on: http://review.whamcloud.com/7841
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-3027 clio: Do not shrink sublock at cancel

Shrinking sublock at ldlm lock cancel time means whoever happened
to attach to this lock just before will reenqueue the wrong lock.

Test-Parameters: envdefinitions=SLOW=yes,ENABLE_QUOTA=yes  clientdistro=el6 serverdistro=el6 clientarch=x86_64  serverarch=x86_64 mdtcount=1  testlist=parallel-scale,parallel-scale-nfsv3,parallel-scale-nfsv4
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Jian Yu <jian.yu@intel.com>
Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Change-Id: I599da98ce364ccf459b5b5f3b72550115972c456
Reviewed-on: http://review.whamcloud.com/7569
Reviewed-by: Patrick Farrell <paf@cray.com>
Tested-by: Hudson
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>

LU-3892 ofd: setattr doesn't need to return attributes

the root cause of the problem is that getattr finds the
object being already destroyed, so we have transno from
preceeding setattr and -ENOENT from getattr and this
confuses OSP.

there is actually no need in attributes in the reply.

Signed-off-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Change-Id: Ib059a443af164d0def58fdfbf41ab545c057cd10
Reviewed-on: http://review.whamcloud.com/7672
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Reviewed-by: Johann Lombardi <johann.lombardi@intel.com>
Reviewed-by: Niu Yawei <yawei.niu@intel.com>

LU-3719 ldiskfs: adjust s_mb_prealloc_table_size correctly

When mb prealloc table item value is not valid, the
s_mb_prealloc_table_size should be adjust accordingly.

ported to SLES11 SP[23]

Lustre-change: http://review.whamcloud.com/7591
Lustre-commit: 6c9f72a1e537826be2b889feba1ff8c5b5db2e77

Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Change-Id: I1803c201743fccf9c177a7098959083f7baf95a0
Reviewed-on: http://review.whamcloud.com/7781
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>

LU-3719 ldiskfs: adjust s_mb_prealloc_table_size correctly

When mb prealloc table item value is not valid, the
s_mb_prealloc_table_size should be adjust accordingly.

Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Change-Id: I17c0d440e4c89cae1780de63ca2b3a0d69de298b
Reviewed-on: http://review.whamcloud.com/7591
Tested-by: Hudson
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Alexander Boyko <alexander_boyko@xyratex.com>
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>

LU-2779 osc: osc_extent_wait() shouldn't be interruptible

Otherwise it will hit the assertion at cl_lock.c:

   cl_lock.c:1967:discard_cb())
ASSERTION( (!(page->cp_type == CPT_CACHEABLE) ||
(!PageWriteback(cl_page_vmpage(env, page)))) ) failed:

This is because in osc_lock_flush() we have to make sure the IO
is finished before discarding the pages.

Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Change-Id: I296ab35d26b861f3c57ddba9f74564aa439aaa0b
Reviewed-on: http://review.whamcloud.com/5419
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-3416 llite: reset writeback index in ll_writepages

Otherwise after one round the writeback index will become beyond
the file size and ->writepages() turns into an empty operation.

Also, a safety guard is added to limit the wait time for grant to
10 minutes(take recovery into consideration) at maximum in the
osc_enter_cache() function. Otherwise, EDQUOT will be returned to
the applications to start sync write.

Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Change-Id: I7d0b90ef0c15f069ec63435703622a327b75ad70
Reviewed-on: http://review.whamcloud.com/7657
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Tested-by: Hudson
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2745 lnet: fix failure handle of create reply

error handler of lnet_create_reply_msg() didn't release lnet_res_lock
if lnet_msg_alloc() failed.
It can be fixed by moving validation check of msg out from lock.

Signed-off-by: Liang Zhen <liang.zhen@intel.com>
Change-Id: I5a25b7e07d24f5c4f165a3266cca6e79dffc1461
Reviewed-on: http://review.whamcloud.com/7659
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Tested-by: Hudson
Reviewed-by: Doug Oucharek <doug.s.oucharek@intel.com>
Reviewed-by: Isaac Huang <he.huang@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-4023 build: wrong type used

Fixed implicit conversion from 'unsigned long long' to 'int'.

Signed-off-by: Dmitry Eremin <dmitry.eremin@intel.com>
Change-Id: I331a8ebe1fb9ef53ebd4fc92603b4100f006ee2b
Reviewed-on: http://review.whamcloud.com/7849
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Alexey Lyashkov <alexey_lyashkov@xyratex.com>

LU-3934 scrub: detect upgraded from 1.8 correctly

The OI scrub should has the capability to detect the case of the MDT
device upgraded from lustre-1.8 and trigger full OI scrub to process
IGIF mapping into OI files.

Originally, we detect that by checking /ROOT/.lustre via initial OI
scrub, but if the MDT is upgraded to lustre-2.1 firstly, and then
continue to lustre-2.4, the such detect mechanism will be failed.
Because lustre-2.1 does not support OI scrub, but will create the
/ROOT/.lustre.

The new detect mechanism will combine the checking OI file name and OI
scrub successfully scan count: if there is old OI file "oi.16" and OI
scrub has never successfully run on the device, then it is regarded as
upgrading case.

Test-Parameters: testlist=sanity-scrub
Signed-off-by: Fan Yong <fan.yong@intel.com>
Change-Id: I5e57aae972e4fe93d3cb9e7cc7ca2e5b95dce4b2
Reviewed-on: http://review.whamcloud.com/7625
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: wangdi <di.wang@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>

LU-3894 mdt: return EXDEV for cross-MDT rename

If target does not exist, and source directory is remote
directory, it should return -EXDEV.

Add test in test_24x to verify it.

Create local directory for test_214, since cross-MDT rename
is not allowed yet.

Signed-off-by: wang di <di.wang@intel.com>
Change-Id: I0d4d8d96249ec23fcea5ab6ac5fce196728e8e73
Reviewed-on: http://review.whamcloud.com/7579
Tested-by: Hudson
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>

LU-3978 nrs: ORR/TRR is using CRR-N fields

A repeated typo is causing ORR/TRR to make use of CRR-N fields;
this completely impairs the ability of ORR/TRR to perform any
sensible scheduling of RPCs, since significant fields used
to enable ORR/TRR scheduling were being overwritten, and
incorrect fields were used when making scheduling decisions.

Signed-off-by: Nikitas Angelinas <nikitas_angelinas@xyratex.com>
Change-Id: I4e30360cd412a37cf14348009c6aedb9843df951
Xyratex-bug-id: MRP-1355
Reviewed-on: http://review.whamcloud.com/7708
Tested-by: Hudson
Reviewed-by: Liang Zhen <liang.zhen@intel.com>
Reviewed-by: Li Xi <pkuelelixi@gmail.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>

LU-3676 llite: error setting max_cache_mb at mount time

The root cause is that when max_cache_mb conf parameter is applied,
the client isn't connected to the OST yet so that sbi->ll_dt_exp is
NULL. However, it's not necessary to shrink the cache memory in this
case so success should be returned.

Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Change-Id: Ib12800812b635207bf4001e2272aadb0ce1e0d6f
Reviewed-on: http://review.whamcloud.com/7194
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>

LU-3273 mdt: Hold med_open_lock before walking med_open_head

Fixed a bug where during replay mdt_mfd_open() calls mdt_handle2mfd()
without acquiring the med_open_lock.
We now take the med_open_lock before traversing med_open_head list.
This bug was noticed during the analysis of LU-3233.

Signed-off-by: Swapnil Pimpale <spimpale@ddn.com>
Change-Id: Ib879f65d41d35f266897e8961dac78e6c4f0d9ec
Reviewed-on: http://review.whamcloud.com/7272
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>

New tag 2.4.1-RC2

Change-Id: I8246d87f0a85b84f45d537cb4feb17b122829ef8

Revert "LU-3027 clio: Do not shrink sublock at cancel"

This uenxpectedly let do LU-7834 whic appears to be old problem,
but happens more frequently with lu-3027 patch in place.
Reverting till the better solutoin is found as the side-effect is worse
than the original problem.

This reverts commit 3781d4465c9fa72120f35084d4eb5edb55a0b66a.

LU-3305 quota: avoid unnecessary dqget/dqput calls

Because of the global lock contention inside the dqget()/dqput(),
we'd try to avoid dqget()/dqput() calls in dquot_initialize().

Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Change-Id: I4335d880f3b912f3de70816a9bd87841605f8907
Reviewed-on: http://review.whamcloud.com/6440
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>

LU-3781 mdt: Atomically update MDT exp_connect_data

don't copy exp_connect_data in mdt_connect_internal()
it restores original LU-1623 behaviour.

Xyratex-bug-id: MRP-1208
Signed-off-by: Andriy Skulysh <Andriy_Skulysh@xyratex.com>
Change-Id: I71ab4d8917bba8ce12fd9af933885796869321e6
Reviewed-on: http://review.whamcloud.com/7391
Tested-by: Hudson
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Niu Yawei <yawei.niu@intel.com>

Tew tag 2.4.1-RC1

Change-Id: I538ebce8c2d7ddf50279cc5a5c90abda9e47362c

LU-3847 kernel: Kernel update [RHEL6.4 2.6.32-358.18.1.el6]

Kernel update [RHEL6.4 2.6.32-358.18.1.el6]

Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Change-Id: I752d3330d4a974bcf2b91392eca449d973cee6c1
Signed-off-by: Jian Yu <jian.yu@intel.com>
Reviewed-on: http://review.whamcloud.com/7516
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

Revert "LU-3544 nfs: writing to new files will return ENOENT"

This commit created a problem lu3765 - crash of 2.1 servers when
exposed to clients with this patch and we don't have a robust
solution in sight for it, so reverting

This reverts commit 77a3c625c733edfaa748d2c7130d5635ee63ce15.

LU-3027 clio: Do not shrink sublock at cancel

Shrinking sublock at ldlm lock cancel time means whoever happened
to attach to this lock just before will reenqueue the wrong lock.

Test-Parameters: envdefinitions=SLOW=yes,ONLY=write_disjoint \
clientdistro=el6 serverdistro=el6 clientarch=x86_64 \
serverarch=x86_64 testlist=parallel-scale

Change-Id: I8f2de683812621fb2f8d761cf2aceebc12868d75
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Jian Yu <jian.yu@intel.com>
Reviewed-on: http://review.whamcloud.com/7481
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Bobi Jam <bobijam@gmail.com>

LU-3063 osc: Check return code for lu_kmem_init

lu_kmem_init can fail and returns has a return code.
Check for this return code in lu_kmem_init.

This issue was found during 2gb VM Racer testing

Signed-off-by: Keith Mannthey <keith.mannthey@intel.com>
Change-Id: I4ca9d5bae212591d5109b2053940161f9e28baeb
Reviewed-on: http://review.whamcloud.com/6514
Tested-by: Hudson
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Reviewed-on: http://review.whamcloud.com/7388

LU-2904 llite: return compatible fsid for statfs

Lustre uses 64-bits inode number to identify object on client side.
When re-export Lustre via NFS, NFS will detect whether support fsid
via statfs(). For the non-support case, it will only recognizes and
packs low 32-bits inode number in nfs handle. Such handle cannot be
used to locate the object properly.
To avoid patch linux kernel, Lustre client should generate fsid and
return it via statfs() to up layer.

To be compatible with old Lustre client (NFS server), the fsid will
be generated from super_block::s_dev.

Signed-off-by: Fan Yong <fan.yong@intel.com>
Change-Id: Id30ff2b2aa5c67c52b09765f50b2afa70cb6d1e3
Reviewed-on: http://review.whamcloud.com/7435
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Hudson
Reviewed-by: Alexey Lyashkov <alexey_lyashkov@xyratex.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-3649 lfsck: release object reference when reset lfsck

When reset lfsck, related lfsck tracing file will be re-created,
but we forgot to release the object reference of the old tracing
file, then trigger LASSERT() when related device to be released.

Signed-off-by: Fan Yong <fan.yong@intel.com>
Change-Id: I657e539183675e00fe43bbbdfde987c3e48dd13b
Reviewed-on: http://review.whamcloud.com/7432
Tested-by: Hudson
Reviewed-by: Patrick Farrell <paf@cray.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>

LU-2547 tests: EXCLUDE recovery-small/24b for ZFS

Exclude test because it fails in the same way as 24a and in theory
should be fixed simultaniously.

Signed-off-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Change-Id: Iea5aaddf41b5765efce22cc7e32234cb1a8e49ae
Reviewed-on: http://review.whamcloud.com/6587
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Hudson
Reviewed-by: Jian Yu <jian.yu@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-on: http://review.whamcloud.com/7424

LU-3620 mdd: missing CLF_RENAME_LAST flag in CL_RENME record

mdd_rename() should use target's attribute to determine add
CLF_RENAME_LAST flag in changelog record.

Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Change-Id: Id9a738f1c98f0c893b125d49965354f44c257e08
Reviewed-on: http://review.whamcloud.com/7095
Tested-by: Hudson
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Fan Yong <fan.yong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Reviewed-on: http://review.whamcloud.com/7384
Reviewed-by: Jian Yu <jian.yu@intel.com>

LU-3180 tests: misc fix in lfsck.sh

- Make sure data flush back before run lfsck;
- Go back to old lfsck behaviour: supress the error of the second run
  of lfsck;
- Fix typo in is_empty_fs();

Test-Parameters: envdefinitions=SLOW=yes testlist=lfsck
Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Change-Id: Ib2c3ab35424e13b1679f8a096f5e24042c7be3a2
Reviewed-on: http://review.whamcloud.com/7387
Tested-by: Hudson
Reviewed-by: Jian Yu <jian.yu@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Emoly Liu <emoly.liu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-3712 lmv: choose right MDT for open by FID

If the open is by FID, LMV should locate MDT by the FID
of the file/directory, instead of its parent.

Signed-off-by: wang di <di.wang@intel.com>
Change-Id: If7815de8cab0cf28b521decccda3c5f0f1aed73c
Reviewed-on: http://review.whamcloud.com/7253
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

Revert "LU-2665 mdc: Keep resend FLocks"

This causes LU-3701

This reverts commit 5e9646ded8a35f1f0b7f3f1eadbd8ab4d927c4a5.

LU-2901 ldlm: fix resource/fid check, use DLDLMRES

In ll_md_blocking_ast() the FID/resource comparison is incorrectly
checking for fid_ver() stored in res_id.name[2] instead of name[1]
changed since http://review.whamcloud.com/2271 (commit 4f91d5161d00)
landed.  This does not impact current clients, since name[2] and
fid_ver() are always zero, but it could cause problems in the future.

In ldlm_cli_enqueue_fini() use ldlm_res_eq() instead of comparing
each of the resource fields separately.

Use DLDLMRES/PLDLMRES when printing resource names everywhere.

Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: Lai Siyao <lai.siyao@intel.com>
Change-Id: I7b45d382d2ee87c1df813ebdd9c7f21029500c1e
Reviewed-on: http://review.whamcloud.com/6592
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Johann Lombardi <johann.lombardi@intel.com>
Reviewed-on: http://review.whamcloud.com/7316
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2901 mdt: duplicate link names in directory

When creating a hard link to a file, the MDT/MDD/OSD code does
not verify whether the target link name already exists in the
directory.  The ZFS ZAP code checks for duplicate entries. The
add_dirent_to_buf() function in ldiskfs only checks entries for
duplicates while it is traversing the leaf block looking for free
space.  Even if it scanned the whole leaf block, this would not
work for non-htree directories since there is no guarantee that
the name is being inserted into the same leaf block.

To fix this, link should check target object doesn't exist as
other creat operations.

Add sanity.sh test_31o with multiple threads racing to link a new
name into the directory, while ensuring that there is a free entry
in the leaf block that is large enough to hold the duplicate name.
This needs to be racy, because otherwise the client VFS will see
the existing name and not send the RPC to the MDS, hiding the bug.

Add DLDLMRES/PLDLMRES macros for printing the whole lock resource
name (including the name hash) in LDLM_DEBUG() messages in a format
similar to DFID/PFID so they can be found in debug logs more easily.

Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: Lai Siyao <lai.siyao@intel.com>
Change-Id: Iaec6098332fe6b0f6a534f8dbecfb14f6f500c1e
Reviewed-on: http://review.whamcloud.com/6591
Tested-by: Hudson
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: Mike Pershin <mike.pershin@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-on: http://review.whamcloud.com/7315
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-3445 utils: make add_param() handle comma-separated values

Current mkfs.lustre supports comma-separated values in "--mgsnode",
"--failnode", etc. options. However, in mount.lustre, the values
are directly passed to mount() as options without prefixing with
Lustre parameter key for each one. This will make the values can
not be recognized by the backend filesystem and cause mount.lustre
fail with -EINVAL error.

This patch improves add_param() to extract values from the comma-
separated string and prefix Lustre parameter key to each value.

Signed-off-by: Jian Yu <jian.yu@intel.com>
Change-Id: I64c5cb6f61277a93f65f69f63050721e0070c450
Reviewed-on: http://review.whamcloud.com/6686
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Reviewed-on: http://review.whamcloud.com/7344

LU-3554 test:  Don't run Sanity 233 on less than 2.4 servers

2.4 clients 2.3 server interop testing showed that sanity.sh
test 233 needs to be skipped for older servers.

Add a check for MDS server code that it is greater than 2.3.64

Signed-off-by: Keith Mannthey <keith.mannthey@intel.com>
Change-Id: I3f96b8017c40a95d246061a8a52db8822c172eca
Reviewed-on: http://review.whamcloud.com/7324
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Minh Diep <minh.diep@intel.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2230 tests: disable replay-dual test 21b with ZFS

This patch disables replay-dual test 21b with ZFS because
semantics of replay_barrier_nosync is different on ZFS:
spa_freeze() syncs all changes. IOW, "nosync" (which is
crucial here) doesn't work.

Test-Parameters: envdefinitions=SLOW=yes,ENABLE_QUOTA=yes \
mdtfilesystemtype=zfs mdsfilesystemtype=zfs ostfilesystemtype=zfs \
clientdistro=el6 serverdistro=el6 clientarch=x86_64 \
serverarch=x86_64 testlist=replay-dual

Signed-off-by: Jian Yu <jian.yu@intel.com>
Change-Id: I271839fa626bfb140bd077d895826c6b4539892f
Reviewed-on: http://review.whamcloud.com/7287
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Li Wei <wei.g.li@intel.com>
Reviewed-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-3326 tests: limit the disk space used by run_*.sh

In run_dd.sh and run_tar.sh, limit the disk space to be used
according to the total free disk space and the client counts.

Test-Parameters: envdefinitions=ENABLE_QUOTA=yes \
clientdistro=el6 serverdistro=el6 clientarch=x86_64 \
serverarch=x86_64 \
clientcount=4 osscount=2 mdscount=2 \
austeroptions=-R failover=true useiscsi=true \
testlist=recovery-mds-scale

Change-Id: I61ea1167906312fdb182b67d7169e95cad1b6d72
Signed-off-by: Hongchao Zhang <hongchao.zhang@intel.com>
Signed-off-by: Jian Yu <jian.yu@intel.com>
Reviewed-on: http://review.whamcloud.com/6376
Tested-by: Hudson
Reviewed-by: Minh Diep <minh.diep@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-3101 mgs: deregister the mount for failed case

in server_start_mgs, the mount of MGS should be put if MGS
failed to start up.

Change-Id: I03c7c3c8e0c578391b637fe0bcdf48da5e318db8
Signed-off-by: Hongchao Zhang <hongchao.zhang@intel.com>
Reviewed-on: http://review.whamcloud.com/6035
Tested-by: Hudson
Reviewed-by: Fan Yong <fan.yong@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>

LU-3744 llite: use 64bits flags in ll_lov_setea()

In ll_lov_setea(), setting MDS_OPEN_HAS_OBJS to an int flags will
result in the flags being overflowed.

Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Change-Id: I7c623785579d782dd7ce03e0ca291d32ad3952f9
Reviewed-on: http://review.whamcloud.com/7312
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Emoly Liu <emoly.liu@intel.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-3264 tests: support failover testing with ZFS

This patch improves test-framework.sh to support failover testing with
ZFS under shared storage configuration.

After formatting a ZFS target, "cachefile=none" property will be set
on the ZFS storage pool so that the pool is not automatically imported
on system startup. And then the pool will be exported so as to leave
the import and export operations handled by mount_facet() and stop().

For SOFT failure mode (unmount and mount), "zpool export …" will be
run on the active server node to export the ZFS storage pool, and
"zpool import -f …" will be run on the failover partner node to import
the pool.

For HARD failure mode (power off and on), since the "cachefile=none"
property has been set on the ZFS storage pool on the active server
node, after the node is power-cycled, the pool will not be imported.
Then on the failover partner node, "zpool import -f …" can be run to
import the pool.

This is a reworked patch with fixes merged in.

Formerly Change-Id: I6b9646cb36f7fa4924af03a055ccb137187843a5
Formerly Reviewed-on: http://review.whamcloud.com/6258
Formerly Change-Id: I63ac141fdefcace99a1cf7c366e15f0067fa2a56
Formerly Reviewed-on: http://review.whamcloud.com/6358
Formerly Change-Id: I83678fd8b67de87f5174017634f108e6ae5b30e0
Formerly Reviewed-on: http://review.whamcloud.com/6408

Test-Parameters: envdefinitions=PTLDEBUG=-1,DEBUG_SIZE=128 mdtfilesystemtype=zfs mdsfilesystemtype=zfs ostfilesystemtype=zfs clientdistro=el6 serverdistro=el6 clientarch=x86_64 serverarch=x86_64 clientcount=4 osscount=2 mdscount=2 austeroptions=-R failover=true useiscsi=true testlist=recovery-double-scale
Signed-off-by: Jian Yu <jian.yu@intel.com>
Signed-off-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Change-Id: I6d2fffede5de7a7d4a6bde3d62dd7c6cb60f63fa
Reviewed-on: http://review.whamcloud.com/6429
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

LU-2200 obdclass: be more careful processing server name

Because whole options line gets passed to exclude processing, don't
search from end of passed in argument to determine fsname at
beginning.

Re-enable conf-sanity/32 but do replace_nids on IB.

Test-Parameters: nettypes=o2ib testlist=conf-sanity
Signed-off-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Change-Id: Ibc43417cb71581198e3516c1e431c0d3f668f1e0
Reviewed-on: http://review.whamcloud.com/6197
Tested-by: Hudson
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>

LU-3553 tests: volatile file test fails interop testing

There is no volatile file support in servers before v2.3.60.
Add a version check and skip sanity test 185 entirely on old servers
to allow interop testing of 2.4+ clients on 2.3 servers.

Signed-off-by: Bob Glossman <bob.glossman@intel.com>
Change-Id: If6aef3a782307b5fe659ffcb8a8f8182f83354bd
Reviewed-on: http://review.whamcloud.com/6887
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Nathaniel Clark <nathaniel.l.clark@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>