LU-130 disable page migration

Disable page migration until we implement a proper ->migratepage handler.

Change-Id: I04eb78c0d04c828eabb3598a3e42ebf51daec3c2
Signed-off-by: Johann Lombardi <johann@whamcloud.com>

LU-457 Enable slab debug for RHEL5/x86_64

Debug patch to turn on slab debugging in the RHEL5 kernel.

Change-Id: I6bcd6f6239b9f5b5953981bcb5fafec426b90005
Signed-off-by: Johann Lombardi <johann@whamcloud.com>

LU-410 Revert LU-15 slow IO with read intense application

This reverts commit 12aae18c0ec495f459a1026ad2e86bbf0232b0fe.
This patch indirectly causes a performance drop with SRP.

Change-Id: If1af51ac8c2d45bc4d5cc5502dc59501678baaec
Signed-off-by: Johann Lombardi <johann@whamcloud.com>

Update e2fsprogs version in changelog

Change-Id: Ib0c6cd31dfec23bf92548028d50655b2520b99d4
Signed-off-by: Johann Lombardi <johann@whamcloud.com>

LU-85 Fix missing bracket

Change-Id: I0c3d16320aab019ef0ef7994c34f6ea5f9ded2a2
Signed-off-by: Johann Lombardi <johann@whamcloud.com>

LU-384 Add space to default PATHS

The update in c81b47a724dd4414b8bd79eb99b9737a70fd5855 broke
configure if --with-linux was not specified, causing build
failures because the .config file could not be found.  Fix this.

Update some ancient documentation that is referenced if the
.config file cannot be found to better describe what is needed
to build Lustre with modern kernel packages (for RHEL at least).

Change-Id: I5b373e3ee3ef42c68257c27a4b194c9e00cf023f
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/946
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-421 include lustre/BUIDING in dist tarball

lustre/BUILDING is not currently included in the dist tarball.

Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: I3c4eed5918aa08df0108aeff134bb9a904f15391
Reviewed-on: http://review.whamcloud.com/948
Tested-by: Hudson
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-422 broken lines need continuations

In commit fb7a499bf47530b9eee162990c8f1815b643d30e there was a sed
introduced that was broken across several lines with no line
continuations:

ver=$(echo $ver |
sed -e 's/^v(.*)/\1/'
-e 's/_RC[0-9].*$//'
-e 's/_/./g')

This is, of course, incorrect.

Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: I7dab960da5ca5fbe7b1852873f72487fd753e415
Reviewed-on: http://review.whamcloud.com/949
Tested-by: Hudson
Reviewed-by: Yang Sheng <ys@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-85 build: update main bug report URL

Update bug reporting URL in autoconf and other documentation
to reflect source of release.

Change-Id: I334aab903123cd0a071ee1b29de4bde5ebc59b0d
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/972
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-409 stop loading the llite_lloop module for >= 2.6.32 kernels

The change of block device interface in kernel 2.6.32-131 caused
the llite_lloop module do not work. Since this is an unsupported
feature, let's stop loading this module from load_modules in
test-framework.sh.

Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: Ic8513dbb02caa93f768a95b9f5798a91f7520ecd
Reviewed-on: http://review.whamcloud.com/954
Tested-by: Hudson
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Yang Sheng <ys@whamcloud.com>

LU-408 escape [[ in autoconf macros

autoconf files use m4 which means [ is meaningful to m4. In order
to escape them, one needs to use [[.

Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: I7601226a0abb8a69e0924c6a8b47b4b0b9feb45a
Reviewed-on: http://review.whamcloud.com/936
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Tested-by: Hudson

Set version to 1.8.6-wc1 to get ready for RC1

Change-Id: I7726fa228359fc67aefedc422d9fe8fbd6280b11
Signed-off-by: Johann Lombardi <johann@whamcloud.com>

LU-402 Update to rhel6.1 kernel 2.6.32-131.2.1.el6

Change-Id: Ib16f8db85b9188f7a9f5982e2914eb647d4b1f07
Signed-off-by: Oleg Drokin <green@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/933
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Tested-by: Johann Lombardi <johann@whamcloud.com>

LU-403 Add support for RHEL5 kernel 2.6.18-238.12.1.el5

Change-Id: Idc6ba8b093d3f2a6e3a13016d7264d5f2937755c
Signed-off-by: Oleg Drokin <green@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/930
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Tested-by: Johann Lombardi <johann@whamcloud.com>

LU-406 Fix compile breakage after adding ocd_max_bytes

Update pack_generic and two wiretest.c files to know about padding
and ocd_max_bytes fields in ocd_connect_data

Signed-off-by: Oleg Drokin <green@whamcloud.com>
Change-Id: I3405b88cf170efa94ee4233f3ccfb86db527ed7c
Reviewed-on: http://review.whamcloud.com/932
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Tested-by: Johann Lombardi <johann@whamcloud.com>

Revert "Set version to 1.8.6-wc1 to get ready for RC1"

The version was updated prematurely and breaks all builds

This reverts commit 6a93cc3b0ce9e5a76487151e5af4c208df5c5c01.

Signed-off-by: Oleg Drokin <green@whamcloud.com>
Change-Id: I6da04b65ba8a013599487e2945988f93d7cac4b1

LU-16 reserve OBD_CONNECT_MAXBYTES connect flag

Reserve connect flag landed to master to prevent conflicts.

Change-Id: I9cc8ebfbfd82083085eea793eec9562d2bfb8bd8
Signed-off-by: Johann Lombardi <johann@whamcloud.com>

Set version to 1.8.6-wc1 to get ready for RC1

Change-Id: Id2da3c16b5c3083b03583406fdd5094318a51229
Signed-off-by: Johann Lombardi <johann@whamcloud.com>

LU-384 land bz22281 on b1_8

To facilitate distinctive naming, we need the work in bugzilla
22281 landed on b1_8.
Add an lbuild option to set the downstream-release.  Normally
this should not be exposed to/through lbuild as this was orig-
inally intended for downstream packages to add their own stamp
to the release string.

Include e4189c54470deff4a98f0b2bb03ac6be7ed85ce9:
    b=24095 don't really need configure --with-linux

    Don't really need to have run configure --with-linux for
    version-tag.pl to work.

    i=wangyb
    i=cliff

Expand the list of debian/* files that should be included in the
dist tarball.
Don't try to [un]patch the local-env.sh.  It should never really
be included in a landed patch.

Include most of 086ff8c46009479bfdac9f570331a6c93b7d52ac:

    b=24094 add kver; update kdir, kobjdir

    Add a new macro to allow the caller to simply define which kernel
    he wants to build for.
    Refactor the kdir and kobjdir to function more properly.
    Remove krequires.  Nothing is using it anymore.

    i=wangyb
    i=cliff

Include c1a61e3e9129d0e2062eed617d922e6dfa9ede5a:

    LU-284 strip .$arch off of kern vers for requires

    RHEL6 appends .arch to the end of the version in the utsrelease.h file.
    The kernel RPM however provides that value with the .arch stripped off,
    so do the same stripping in the lustre-modules RPM for it's Requires:.

Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
    Change-Id: Ib6ff32c006f184edceb5d667b649bb106d503faf
    Reviewed-on: http://review.whamcloud.com/502
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Tested-by: Hudson
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: I82f4f9a556fd1d1fc51ef8930e2017798ff88706
Reviewed-on: http://review.whamcloud.com/884
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Tested-by: Hudson

LU-400 don't unroll empty commits into patches

The patch unrolling tool, build/extract_patches, currently will
create an empty dpatch if a commit is empty. This causes dpatch
some heartburn, so just skip empty commits.

Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: Ice0e5fad3c1dd01b507838a8b449536d016b8ffb
Reviewed-on: http://review.whamcloud.com/908
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Tested-by: Hudson

Remove changelog entry for LU-179

... this was already landed upstream (in Oracle's 1.8.6)

Change-Id: I29f23b9a330847f91ba0de3883396de67966896c
Signed-off-by: Johann Lombardi <johann@whamcloud.com>

Add missing changelog entries

Change-Id: I15dbcc4eef08e2afbfa223ae87cf31b7654a14e9
Signed-off-by: Johann Lombardi <johann@whamcloud.com>

Add OFED 1.5.3.1 to supported version and update recommended e2fsprogs version

Change-Id: Ie2b5daf4df98bbcd9bb84b558948269378b40df0
Signed-off-by: Johann Lombardi <johann@whamcloud.com>

LU-85 Update some URLs and text to reflect our releases.

The text and URLs in the debian packaging metadata were old
(circa Sun Microsystems), so update them and take the opportunity
to inform about the source of the release being installed.

Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: I23e87c0f093e4da208f4b66873ea9849fa62cc25
Reviewed-on: http://review.whamcloud.com/906
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Hudson

LU-376 Cleanup wirecheck

Drop unused ones and add 'MDS_DIR_END_OFF' check.

Change-Id: I813a2963d9ce858386f3cfff8581536a2e46d981
Signed-off-by: nasf <yong.fan@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/891
Tested-by: Hudson
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-364 remove . - in node name

Replace . and - with _ in the node name, so it can be identified
as a variable in bash.

Change-Id: I7b5de328728c0b2433e2cea73d3520648eb9fdf2
Signed-off-by: sarah <sarah@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/885
Tested-by: Hudson
Reviewed-by: Chris Gearing <chris@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-303 fix ofed download for bugfix releases

On the OFED download server both w.x.y and w.x.y.z releases are
stored in the w.x.y dir, so handle that properly for the w.x.y.z case.

Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: I7e638571a31df215517d71e5997e7d0c6411bb90
Reviewed-on: http://review.whamcloud.com/864
Tested-by: Hudson
Reviewed-by: Yang Sheng <ys@whamcloud.com>
Reviewed-by: Michael MacDonald <mjmac@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

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.

In addition, "set_default_debug_nodes()" has a defect to check
whether "$nodes" contains the local node or not. The patch contains
a fix for that.

Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: Iae49497601d09fef38968360d8543cc753009e85
Reviewed-on: http://review.whamcloud.com/871
Tested-by: Hudson
Reviewed-by: Brian J. Murrell <brian@whamcloud.com>
Reviewed-by: Chris Gearing <chris@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-388 modify kernel-ib spec file for openib RPM

The kernel-ib RPM from OFED 1.5.3.1 conflicts with RHEL5's
openib package.  Consider RHEL5's userspace as canonical and
Remove the conflicting portions from the kernel-ib RPM.

Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: I8791ffabef7fc498486ff5b5dda50f49377e8466
Reviewed-on: http://review.whamcloud.com/896
Tested-by: Hudson
Tested-by: Yu Jian <yujian@whamcloud.com>
Reviewed-by: Michael MacDonald <mjmac@whamcloud.com>
Reviewed-by: Yang Sheng <ys@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-303 fix OFED bugfix release download location

On the OFED download server both w.x.y and w.x.y.z releases are
stored in the w.x.y dir, so handle that properly for the w.x.y.z case.

Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: I28d63d8a64d50665e905b3f8971bccfe6bad0fe9
Reviewed-on: http://review.whamcloud.com/895
Tested-by: Hudson
Reviewed-by: Yang Sheng <ys@whamcloud.com>
Reviewed-by: Michael MacDonald <mjmac@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-363 don't cache BUILD for reuse

Since future builds don't actually need the %_topdir/BUILD subdir,
and it's huge, don't bother caching it.

Increment the toplevel BUILD_GEN to force a rebuild of all of the
kernels.

Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: I6a7a523fdaed2073d719f44a24b556bc4595d4ec
Reviewed-on: http://review.whamcloud.com/863
Tested-by: Hudson
Reviewed-by: Yang Sheng <ys@whamcloud.com>
Reviewed-by: Chris Gearing <chris@whamcloud.com>
Reviewed-by: Michael MacDonald <mjmac@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-376 Positive LL_DIR_END_OFF to indicate the tail of dir hash/offset

1) Keep 'MDS_DIR_END_OFF' unchanged (0xfffffffffffffffeULL) to simplify
   dir hash/offset related interoperability issues.
2) Introduce positive "LL_DIR_END_OFF" (0x7fffffffffffffffULL) on client
   to indicate the tail of dir hash/offset for up layer callers, like
   llseek(), readdir(), and so on.
3) Support 1.8 client to talk with old 2.0 server with 32bit hash.

Change-Id: I126ddb170b9ee24d1ae0610ac6343c9b0f5e4c70
Signed-off-by: nasf <yong.fan@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/887
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

Bump version to 1.8.5.56 and add changelog section for 1.8.6-wc

Change-Id: I0f99f1abb0c7fb87f9fecc625639914075385f8a
Signed-off-by: Johann Lombardi <johann@whamcloud.com>

LU-196 set $PTLDEBUG, $SUBSYSTEM and $DEBUG_SIZE values on every node

The current set_default_debug_nodes() could not pass the values of
$PTLDEBUG, $SUBSYSTEM and $DEBUG_SIZE to the remote nodes while they
are specified from the command line on the local node. This patch is
to fix this issue.

Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: I85922c67e7c68a5ea3e9ba7fc1cbf6516907c285
Reviewed-on: http://review.whamcloud.com/593
Tested-by: Hudson
Reviewed-by: Chris Gearing <chris@whamcloud.com>
Reviewed-by: Yang Sheng <ys@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-136 change "force_over_16tb" mount option to "force_over_24tb"

Change the "force_over_16tb" mount option to "force_over_24tb" and
rename the ext4-force_over_16tb-*.patch to ext4-force_over_24tb-*.patch
after testing and validating the 24TB LUN.

Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: Ia1731216a0c5fd2e0de4ec46c4cbd32bcd7c691d
Reviewed-on: http://review.whamcloud.com/589
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>

LU-228 Enabled ext4-based ldiskfs by default on RHEL5

Change-Id: Ic4f4a94d86091acab1a77a0cbf1d9a24c715e2c0
Signed-off-by: Yang Sheng <ys@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/574
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>

LU-170 oscc_grow_count will never grow

We are using req::rq_async_args.space[0] to store original value of
oscc_grow_count, and using req::rq_async_args.pointer_arg[0] to store
oscc, however, ptlrpc_async_args is a union, which means
req::rq_async_args.space[0] will always be overwritten by a ossc (pointer),
and osc_interpret_create will always get true on this condition
"if (diff < (int) req->rq_async_args.space[0])" and reset
oscc_grow_count to OST_MIN_PRECREATE and set OSCC_FLAG_LOW.

Because it's very unsafe to use raw scratchpad directly, I also cleaned
up all using of raw scratchpad in this patch.

Change-Id: I431a37521a41c5dfbe10ebca9efac157985add51
Signed-off-by: Liang Zhen <liang@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/588
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-264 ext4 MMP update

Integrate the ext4 MMP update into ldiskfs

Signed-off-by: Niu Yawei <niu@whamcloud.com>
Change-Id: Iec06c5d73c5ae426458c1141dcd38d96b1a2b7f4
Reviewed-on: http://review.whamcloud.com/561
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Tested-by: Maloo <whamcloud.maloo@gmail.com>

LU-337 Fix alloc flags in alloc_qinfo()

CFS_ALLOC_IO allows a hang when the alloc_qinfo() results
in a cache_alloc_refill, and kicks off the ldlm shrinker.
The shrinker may hang in sync_page() on the same page on
which we already hold the lock.

Change-Id: I3d9887ebc2909d38b4abb2e8d7715913c9c2cc40
Signed-off-by: Christopher J. Morrone <morrone2@llnl.gov>
Reviewed-on: http://review.whamcloud.com/582
Reviewed-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Tested-by: Hudson

LU-295 Remove 'lustre_rsync-test' from DEFAULT_SUITES

The 'lustre_rsync-test' is not relevant (nor does it exist) to b1_8
so we should remove this test from the DEFAULT_SUITES list.

Signed-off-by: Chris Horn <hornc@cray.com>
Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: I988d3124d6fae06afca013acfa94d32c6fee24d7
Reviewed-on: http://review.whamcloud.com/515
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Tested-by: Hudson

LU-234 OOM killer causes node hang.

b=18213 Handle the signal to avoid process hang.

Previous patch may cause a memleak. And also release the request when
ldlm_cli_enqueue() failed.

Signed-off-by: Yang Sheng <ys@whamcloud.com>
Change-Id: I1c22e2b2027f61af257ce031cfb7b6cbe08b8141
Reviewed-on: http://review.whamcloud.com/457
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-by: Fan Yong <yong.fan@whamcloud.com>

LU-216 Protect extent tree during fsfilt_ldiskfs_ext_walk_space()

Guarantee the validity of the extent path passed in
ldiskfs_ext_new_extent_cb().

Change-Id: I3f80e1e935d8ce02233cbdbf5d0ebb60dc50d8b9
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/573
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-339 use new aops for kernel >= RHEL5.4

for RHEL5 kernel, from >= 5.4 both old and new aops interface are
supported, and a flag MS_HAS_NEW_AOPS is used to mark wheather a
filesystem supports it. Enable it for lustre by default, and this
will increase single thread IO perf > 5% due to kernel
implementation.

Signed-off-by: Lai Siyao <laisiyao@whamcloud.com>
Change-Id: I8f0bfb1c4eea39275cb22c52e30310dcad41db14
Reviewed-on: http://review.whamcloud.com/570
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-by: Yang Sheng <ys@whamcloud.com>

LU-200 1.8.6 client supports 64-bit dir name hash under interoperability mode

1.8.6 client supports 64-bit dir name hash under interoperability mode.

Signed-off-by: nasf <yong.fan@whamcloud.com>
Change-Id: If8130cfe809424502f40ff32dcb89deb8b82d720
Reviewed-on: http://review.whamcloud.com/410
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-216 Restore WALK_SPACE_HAS_DATA_SEM check.

Change-Id: Id0307a85890914a9d1695b0faa82fb89f278e1f4
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/447
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-by: Niu Yawei <niu@whamcloud.com>

LU-286 racer: general protection fault.

Fixed previous patch hasn't cleanup appropriate may cause
invalid reference.

Change-Id: Ib9b6375c9a7bc57ee5ad17da514848ce8ca87acd
Signed-off-by: Yang Sheng <ys@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/506
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-265 Improve autogen

Simplify autogen.sh and tell the user if an auto* command failed.

Change-Id: Ia764fedffe182dd431576eb1d047b97d6813ea7b
Signed-off-by: Bernd Schubert <bschubert@ddn.com>
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/492
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Tested-by: Hudson

LU-296 Add missing code from auster/yaml port to b1_8

Code related to defining and creating the $TF_FAIL file was missed in
the port of auster and yaml to b1_8. This patch adds the missing
code. Namely:

- Define TF_FAIL in acceptance-small.sh
- add 'touch $TF_FAIL' to the test-framework.sh ERR signal trap

Additionally, the file test for $TF_FAIL variable should be quoted in
acceptance-small.sh and auster.sh. Otherwise, if this variable is not
defined, bash will spit out something like

acceptance-small.sh: line 123: [: too many arguments

Bash variables should generally be quoted whenever they are referenced.
(see http://tldp.org/LDP/abs/html/quotingvar.html).

Signed-off-by: Chris Horn <hornc@cray.com>
Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: I2071c859d00b8cd8bd423bf9f626b44352456191
Reviewed-on: http://review.whamcloud.com/516
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Tested-by: Hudson

LU-240 test_171() races with background multiop

Change-Id: Ibdcc584218708bd29724f3f71804a3ebf5b1d454
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/520
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-228 kernel update to rhel5 2.6.18-238.9.1

Clear up the kernel support information.

Signed-off-by: Yang Sheng <ys@whamcloud.com>
Change-Id: I6b4364eb8d22629a1caee202e4b77c7f197b5b01
Reviewed-on: http://review.whamcloud.com/539
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-231 sanity_test_170: FAIL: expected 248 bad lines, but got 249

if the length of malformed line is less than sizeof(struct ptldebug_header),
it's not counted into bad lines if this line is at the end of log file.

Signed-off-by: Lai Siyao <laisiyao@whamcloud.com>
Change-Id: Ib2cb1a81f2b8d00c855f11d232f7643e66cebeb0
Reviewed-on: http://review.whamcloud.com/526
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-274 Update LVB from disk when glimpse callback return error

  Client ll_glimpse_callback() could fail to get inode if the inode is
  already been cleared, and this glimpse callback will fail for
  -ELDLM_NO_LOCK_DATA, so server should update LVB from disk
  (in filter_intent_policy()) when it received such error from client.

Issue:
Signed-off-by: Niu Yawei <niu@whamcloud.com>
Change-Id: I29b2fe4b35114359b0ad62106de48f5f2f696636
Reviewed-on: http://review.whamcloud.com/507
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-228 kernel update to rhel5.6.

build ok on 2.6.18-238.9.1.el5.

Signed-off-by: Yang Sheng <ys@whamcloud.com>
Change-Id: Id66c5fc89b233ef59e851b78ca52f4ea2159c9d8
Reviewed-on: http://review.whamcloud.com/455
Tested-by: Hudson
Reviewed-by: Brian J. Murrell <brian@whamcloud.com>
Reviewed-by: Michael MacDonald <mjmac@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-232 do not use bash variable to pass test setup state between sub-tests

Port the patch from bug 22842 attachment 30538 to b1_8 branch:
1) fix sanity tests 42, 101 and 102
2) fix PATH variable to search utilities from $LUSTRE/tests dir first

Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: Idb58789efc1a1a0a50b54db5cb1f6bad5b5f8daa
Reviewed-on: http://review.whamcloud.com/467
Tested-by: Hudson
Reviewed-by: Chris Gearing <chris@whamcloud.com>
Reviewed-by: Brian J. Murrell <brian@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-118 Workaround for race causing assertion failure in clear_inode()

This patch is just a *workaround* to avoid a race found in recent
kernel causing the following assertion in clear_inode():
BUG_ON(inode->i_data.nrpages).
The race can still occur when truncate_inode_pages() & clear_inode()
are called directly by the VFS, but the problem should be harder to
reproduce.
A more complete fix is considered for 2.x.

Change-Id: If0df4b1cba00de7cf43f1b3510d54d35525c5e14
Signed-off-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/475
Tested-by: Hudson
Reviewed-by: Jinshan Xiong <jay@whamcloud.com>
Reviewed-by: Andrew Perepechko <andrew_perepechko@xyratex.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-262 port the changes of llverdev.c and llverfs.c from master to b1_8

The main changes are:
bug 24017 make llverfs more robust against errors during testing
bug 20492 allow llverdev/llverfs to handle media errors better

Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: Ib7df0c98304d8fa56f3ad69127617147c2987b8d
Reviewed-on: http://review.whamcloud.com/487
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>

LU-235 fix sanity test 36g to work with remote OSS

Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: Ia4e0b01d002783bc8d4f38e7c1e6b1ac753144d6
Reviewed-on: http://review.whamcloud.com/473
Tested-by: Hudson
Reviewed-by: Chris Gearing <chris@whamcloud.com>
Reviewed-by: Brian J. Murrell <brian@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-15 slow IO with read-intense application

Align the readahead extent by 1M after when it is trimed by ra_max_pages.

signed-off-by: Wang Di <di.wang@whamcloud.com>
Change-Id: I0359a39cef678b05e48617910081e9f27826ed22
Reviewed-on: http://review.whamcloud.com/437
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Reviewed-by: Jinshan Xiong <jay@whamcloud.com>
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-202 fix the test_23 of sanity-quota.sh

test_23 uses a constant value(120000 = ~120M) to check if the system has
enough space to finish the test, so in the large OSTCOUNT testing
environment, the test will fail for short of disk space.

Changes of this patch are:
- Use a reasonable value to check free space instead of magic 120000;
- Specify the stripe index of test file for test_9/10 and test_23;
- Check free space over all OSTs, and select the ost which has enough
  space to do test. (test_9/10/23)

Issue:
Signed-off-by: Niu Yawei <niu@whamcloud.com>
Change-Id: I2d9059d23cf53fee0244a226622b0e9adb800599
Reviewed-on: http://review.whamcloud.com/454
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-230 fix yml_entities() to create MDS entity

There is only one MDS on b1_8.

Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: I4ab3f0ad27ab76dd77ef9cc153b932b315ea9c4d
Reviewed-on: http://review.whamcloud.com/464
Reviewed-by: Bobi Jam <bobijam@whamcloud.com>
Tested-by: Hudson
Reviewed-by: Chris Gearing <chris@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-203 pass client mount options to zconf_mount_clients() correctly

Quote the third argument and pass it to zconf_mount_clients().

Signed-off-by: Yu Jian <yujian@whamcloud.com>
Change-Id: I86c2aea284178ca561c809f439faf9fd01d6e1c6
Reviewed-on: http://review.whamcloud.com/461
Tested-by: Hudson
Reviewed-by: Brian J. Murrell <brian@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-by: Chris Gearing <chris@whamcloud.com>

LU-15 slow IO with read intense application

Shrink file_max_cache_size to alleviate the memory pressure of OST.

Change-Id: I6dc26ea88f8fc28bcbc94cf7d1e6f2bdfc866ea7
signed-off-by: Wang Di <di.wang@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/445
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>

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

i=wangdi
i=johann

Change-Id: If44d996f8aa6268bf9e4afd2b1dbfdd8f4460a8c
Signed-off-by: Yang Sheng <ys@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/441
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-123 Port yaml and auster to b1_8

Changes to add the yaml data logging from the 2.0 branch in the 1.8 branch, this
patch was created by applying the 2.0 yml patch to 1.8 and then resolving the issues.

This is to say that the larger changes have been taken to be correct because of their
acceptance in the 2.0 master.

This patch also contains the addition of the auster.sh script which allows for the logging
of test results to the maloo database. The actual logging is carried out by the
maloo_upload.sh script.

For instructions on how to use auster use auster -?, for instructions on maloo refer to
the whamcloud wiki [where I will begin to place a resource of information!]

Change-Id: I602a3534f17544d857aa0a9f9f82d2873fb73a39
Signed-off-by: Chris Gearing <chris@whamcloud.com>
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/421
Tested-by: Hudson
Reviewed-by: Yu Jian <yujian@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-179 Should let PF_MEMALLOC cover ptlrpc_register_bulk()

Under memory pressure, the rpc memory allocation in bulk registration on the memory
release path also needs be covered.

Change-Id: Ie4986eb48d1e5b99b1756a0ddaa11fe0beb37462
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/439
Reviewed-by: Liang Zhen <liang@whamcloud.com>
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-206 Memory corruption due to MMP (1.8.5)

Change-Id: Ibc0c3ea76eb94c7a64f7e17711e9fe4b954cc37b
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: http://review.whamcloud.com/433
Tested-by: Hudson
Reviewed-by: Oleg Drokin <green@whamcloud.com>

LU-129 Skip quotacheck on administratively disabled OSTs

- skip quotacheck on administratively disabled OSTs.

- introduce OBD_NOTIFY_DEACTIVATE & OBD_NOTIFY_ACTIVATE event, thus lov is now aware of
  administratively enable/disable import event, and is able to udpate the 'ltd_activate'
  accordingly.

Issue: LU-129
Signed-off-by: Niu Yawei <niu@whamcloud.com>
Change-Id: Id3d2c9e1e035086e653ab500459912f7960976c8
Reviewed-on: http://review.whamcloud.com/351
Tested-by: Hudson
Reviewed-by: Lai Siyao <laisiyao@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-62 rhel6 support

So that RHEL6 client can be built.

Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: Ie5ff1249aee8725625901d1dad242a8f8c48765a
Reviewed-on: http://review.whamcloud.com/385
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-177 jenkins build enhancements

From fd0bd149e66cbc06a17c372e159f9bfba024cc97:
  b=22281 refactor packaging and versioning

  Only the --enable-dist configure flag.

  i=wangyb
  i=mjmac

From ae4652c1ac831912e186d10214b266b994d7a7e2:
  b=24090 distro and target autodetection

  Automatically detect the target and distro.  It's in fact a little
  crazy that these are two separate options.  They should be merged.

  i=cliff
  i=minh

From 63520d0a8089ba4929fc6767442e85bde4a1ccad:
  b=24091 find_linux_rpms utility

  Write a utility to find the the kernel "devel" package needed.
  Restructure some code to facilitate reuse of existing code to
  use in the utility.

  i=cliff
  i=minh

From 2941269c348cfed9c22d02c45eb00687339adaaa:
  LU-120 the perfect storm

  There is a corner case in the debian patching/unpatching process
  where a patch converts the use of a given file to a file.in that is
  to be substed.  The problem is that between the configure and the
  make debs, a file, which a debian patch wants to remove already
  exists (it was created by the configure) and dpkg-buildpackage's
  unpatch step fails becuase it wants to try to create a file which
  configure already created.

Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
  Change-Id: I185e496b82687d21ea3d744e939536c106b702f9
  Reviewed-on: http://review.whamcloud.com/308
Tested-by: Hudson
Reviewed-by: Robert Read <rread@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
From 98bf9c9e3ecaa6736dcb066d0c76e20610fa9725
  LU-177 lbuild autonomy fixes

  Fixes to allow lbuild to work without uneeded help from it's caller:
  - CentOS is a "rhel" distro
  Direct output to the output FD.

Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
  Change-Id: I714b90b53b14ac4ef4cd6bfe638a029656a151b0
  Reviewed-on: http://review.whamcloud.com/361
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Tested-by: Hudson
Reviewed-by: Michael MacDonald <mjmac@whamcloud.com>
From 8fd9914d125989b924fc6c0e1d50dfd193a73438:
  LU-73 lbuild and friends for jenkins

  Add the portions of this patch needed for lbuild to work in
  our matrix-building jenkins environment.
  Included is a re-org of the RHEL5 lbuild code to maximize code
  reuse and minimze code replication for future RHEL6 landing.

Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Reviewed-by: Michael MacDonald <mjmac@whamcloud.com>
Reviewed-by: Robert Read <rread@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: I6b4c668d16b302f62a4d5afb9c84a051c260af06
Reviewed-on: http://review.whamcloud.com/402
Tested-by: Hudson
Reviewed-by: Michael MacDonald <mjmac@whamcloud.com>
Reviewed-by: Yang Sheng <ys@whamcloud.com>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-73 b=22375 RHEL6 kernel support for 1.8

rhel6 patchless client support:
* different shrinker function type
* blk_queue_max_sectors/segments renamed
* removed obsolete function bio_hw_segments

signed-off-by: Lai Siyao <laisiyao@whamcloud.com>
Change-Id: Ia96cbfafc6213a86b21254aad88d7807256fd0d9
Reviewed-on: http://review.whamcloud.com/282
Tested-by: Hudson
Reviewed-by: Johann Lombardi <johann@whamcloud.com>

LU-57 Allow OSTs to be created with no primary node

Add a --servicenode parameter for mkfs.lustre to treat all service
nodes equally.

Bugzilla: 19064
Change-Id: I7851cb7781cdaf59d2cf18690d2da35121432cc1
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>

b=23935 append truncate race

i=Vladimir.Saveliev
i=Oleg.Drokin

Original patch from Vitaly Fertman

b=21847 obdfilter-survey: Syntax error in some locales

i=Vladimir.Saveliev

b=21501 Properly cleanup flock lock on disconnect

Properly wakeup flock waiters on eviction.
Destroyed lock for flock completion ast is not an error, return success
to avoid double lock decref.

i=rread
i=adilger

b=24437 revoke open lock for executable files if needed

When a normal lustre client open write/exec a file, the open lock on that file needs to be revoked
in case an NFSD lustre client still holds it.

i=green@whamcloud.com
i=vladimir.saveliev@oracle.com

b=22729 Remove LPSZ & LPSSZ

Code cleanup patch for 1.8 which removes the use of LPSZ/LPSSZ to
improve the build portability.

Issue: LU-54
Change-Id: I04d8446564bf36d0806ff3ee9abf8950a33eb170
Signed-off-by: Niu Yawei <niu@whamcloud.com>

LU-53 run autogen if a Makefile.am is patched

Signed-off-by: Brian J. Murrell <brian@whamcloud.com>
Change-Id: I40eb3cecff64a62c2eca212ed19263128ad4b50d

b=21137 Sles11 with 1.8 is slower than 1.6 sles10 for O_DIRECT single file IOR writes

Fix ptlrpc_main() condition to start service threads correctly.

i=adilger
i=vitaly.fertman

b=23049 t-f do_node() VERBOSE fix

i=Vladimir.Saveliev

Cleared white space warning from file.  Deleted
blank line from EOF.

b=24479 LU-92 files and dirs missing in dist tarball

Some files and dirs are missing in the "dist" tarball.

Change-Id: Id9fea606549174cac2ea1f7239d75f2fc63c144c
Signed-off-by: Brian J. Murrell <brian@whamcloud.com>

Updated version for the next incremental build.

Corrected a trailing white space error.

b=19494 "lfs find" hangs when searching for an OST index

i=Emoly.Liu
i=Johann.Lombardi <johann@whamcloud.com>

- new test_88 "lfs find identifies the missing striped file segments"
- exit_status () egrep pattern fix

b=24194 increase reseed count to mitigate inconsistence in OST allocation

in alloc_rr, "LOV_CREATE_RESEED_MULT" and "LOV_CREATE_RESEED_MIN" is increased
to mitigate the inconsistence in OST allocation.

i=johann
i=andreas.dilger

b=24451 racer test cleanup

i=Vladimir.Saveliev
i=Andrew.Perepechko <Andrew_Perepechko@xyratex.com>

- modify racer/racer.sh to wait the process killed,
  exit 1 if the process are still existing;
- remove runracer;

b=19649 sanity test_77j fix

i=Johann.Lombardi <johann@whamcloud.com>

b=24426 add ERRLOG suffix to not ovewrite the lustre logs

i=Vladimir.Saveliev

b=24420 avoid an LASSERT on recovery

o=Diego.Moreno
i=tappro
i=vitaly

b=24375 Fix a race between completion and enqueue

o=green
i=vitaly
i=vs

ldlm_enqueue_tail does not obtain proper lockng when checking lock mode to see
if the lock is granted, so there is a window where ldlm_handle_completion_ast
can update lvb with correct data, but beforeit has a chance to update the lock
mode, the ldlm_enqueue_tail will check the lock mode and since the lock is not
granted yet, it will overwrite correct lvb with stale value from enqueue time.

b=24050 fix lustre_start to start server targets in the order of MGS->MDT->OST(s)

i=vitaly.fertman

b=24426 run_one(): run error() once

there is no reason to run error() (and lctl dk thereby) more than once
second lctl dk overwrites the most important logs obtained on first lctl dk

i=grev

b=23787 Modified struct lprocfs_percpu to be C99 compliant.

i=yangsheng

b=24432 mount_lustre.c/parse_options() fix

o=Chris.Horn <hornc@cray.com>
i=Elena.Gryaznova

to differentiate between 'force*' and 'force'

b=22168 write-append-truncate: retry write when receives EINTR.

r=jian.yu
r=vitaly.fertman

b=22984 change all references to tune.ldiskfs in lustre to tunefs.ldiskfs

o=girish
i=adilger
i=johann

b=21135 calculate Use% for "lfs df" the same way as standard "df"

o=andreas
i=girish
i=simmonsja

b=19944 adjust debug size to be -gt num_possible_cpus()

a=Landen <zhiyong.tian@oracle.com>
i=Elena.Gryaznova

b=23670 exit_status () fix

i=Vladimir.Saveliev

b=23430 fix sanity-quota test 14a to write file in O_DIRECT mode

o=andrew.perepechko
i=johann.lombardi
i=landen