ceph-qa-suite.git
4 years agoMerge pull request #1310 from ceph/wip-18113 master
Samuel Just [Wed, 14 Dec 2016 23:57:24 +0000 (15:57 -0800)]
Merge pull request #1310 from ceph/wip-18113

test: Update for new error message when doing scrub with deep-scrub e…

Reviewed-by: Samuel Just <sjust@redhat.com>
4 years agoRevert "tasks/workunit.py: depth 1 clone"
Sage Weil [Wed, 14 Dec 2016 17:28:23 +0000 (12:28 -0500)]
Revert "tasks/workunit.py: depth 1 clone"

This reverts commit 4faf77a649cb3f8ddf497ca81937b3dbf63a18dc.

4 years agoMerge pull request #1307 from ceph/wip-workunit-repo
Zack Cerza [Wed, 14 Dec 2016 17:21:05 +0000 (10:21 -0700)]
Merge pull request #1307 from ceph/wip-workunit-repo

tasks/workunit: use correct git mirror url

4 years agotasks/workunit.py: depth 1 clone 1307/head
Sage Weil [Wed, 14 Dec 2016 17:19:44 +0000 (12:19 -0500)]
tasks/workunit.py: depth 1 clone

Signed-off-by: Sage Weil <sage@redhat.com>
4 years agotasks/workunit: remove kludge to use git.ceph.com
Sage Weil [Wed, 14 Dec 2016 17:18:29 +0000 (12:18 -0500)]
tasks/workunit: remove kludge to use git.ceph.com

This was hard-coded to ceph.git (almost) and breaks when
you specify --ceph-repo.  Remove it entirely.  We'll see if
github.com is better at handling our load than it used to
be!

Signed-off-by: Sage Weil <sage@redhat.com>
4 years agoMerge pull request #1309 from ceph/wip-16800-kefu
Sage Weil [Wed, 14 Dec 2016 15:56:25 +0000 (09:56 -0600)]
Merge pull request #1309 from ceph/wip-16800-kefu

tasks/ceph: restore context of osd mount path before mkfs

4 years agoMerge pull request #1156 from ceph/wip-kclient-python
John Spray [Wed, 14 Dec 2016 14:09:13 +0000 (14:09 +0000)]
Merge pull request #1156 from ceph/wip-kclient-python

Enable cephfs python tests on kclient

4 years agotasks/ceph: restore context of osd mount path before mkfs 1309/head
Kefu Chai [Fri, 9 Dec 2016 18:36:52 +0000 (02:36 +0800)]
tasks/ceph: restore context of osd mount path before mkfs

all newly created files and directories under the mount dir inherit the
SELinux type of their parent directory. so we need to set it before
mkfs.

Fixes: http://tracker.ceph.com/issues/16800
Signed-off-by: Kefu Chai <kchai@redhat.com>
4 years agotasks/cephfs: set norequire_active_mds on kclient 1156/head
John Spray [Tue, 13 Dec 2016 01:05:25 +0000 (01:05 +0000)]
tasks/cephfs: set norequire_active_mds on kclient

The kernel client's cluster availability test is
more primitive than the fuse client, so we need
to switch it off to avoid client mounts failing
while MDSs are still coming up.

Fixes: http://tracker.ceph.com/issues/18161
Signed-off-by: John Spray <john.spray@redhat.com>
4 years agotest: Update for new error message when doing scrub with deep-scrub errors wip-18113 1310/head
David Zafman [Mon, 12 Dec 2016 23:53:25 +0000 (15:53 -0800)]
test: Update for new error message when doing scrub with deep-scrub errors

Signed-off-by: David Zafman <dzafman@redhat.com>
4 years agoMerge pull request #1304 from vshankar/wip-15028
Jason Dillaman [Mon, 12 Dec 2016 21:29:54 +0000 (16:29 -0500)]
Merge pull request #1304 from vshankar/wip-15028

cram: support fetching from sha1 branch, tag, commit hash

Reviewed-by: Jason Dillaman <dillaman@redhat.com>
4 years agoMerge pull request #1305 from ceph/wip-h-j-x
Sage Weil [Mon, 12 Dec 2016 17:37:30 +0000 (11:37 -0600)]
Merge pull request #1305 from ceph/wip-h-j-x

upgrade/hammer-jewel-x: whitelist quota

4 years agoupgrade/hammer-jewel-x: whitelist quota 1305/head
Sage Weil [Mon, 12 Dec 2016 17:29:09 +0000 (12:29 -0500)]
upgrade/hammer-jewel-x: whitelist quota

This is from teh librados api test.

Signed-off-by: Sage Weil <sage@redhat.com>
4 years agocram: support fetching from sha1 branch, tag, commit hash 1304/head
Venky Shankar [Fri, 9 Dec 2016 09:41:49 +0000 (15:11 +0530)]
cram: support fetching from sha1 branch, tag, commit hash

Signed-off-by: Venky Shankar <vshankar@redhat.com>
4 years agoMerge remote-tracking branch 'upstream/pull/1302/head'
Patrick Donnelly [Fri, 9 Dec 2016 16:09:02 +0000 (11:09 -0500)]
Merge remote-tracking branch 'upstream/pull/1302/head'

* upstream/pull/1302/head:
  fsx-mpi: fsx-mpi expects a file name not a dir
  fsx-mpi: remove chef task
  Fix mpi calls

4 years agofsx-mpi: fsx-mpi expects a file name not a dir 1302/head
Patrick Donnelly [Tue, 6 Dec 2016 21:51:38 +0000 (16:51 -0500)]
fsx-mpi: fsx-mpi expects a file name not a dir

The test still fails even after being enabled:

2016-12-07T18:00:44.337 INFO:teuthology.orchestra.run.mira105:Running: 'mpiexec -f /home/ubuntu/cephtest/mpi-hosts -wdir /home/ubuntu/cephtest/gmnt sudo /home/ubuntu/cephtest/fsx-mpi -o 1MB -N 50000 -p 10000 -l 1048576 /home/ubuntu/cephtest/gmnt/test'
2016-12-07T18:00:44.486 INFO:teuthology.orchestra.run.mira105.stderr:Warning: Permanently added '172.21.8.122' (ECDSA) to the list of known hosts.
2016-12-07T18:00:44.571 INFO:teuthology.orchestra.run.mira105.stdout:skipping zero size read
2016-12-07T18:00:44.591 INFO:teuthology.orchestra.run.mira105.stdout:truncating to largest ever: 0x7cccb
2016-12-07T18:00:44.606 INFO:teuthology.orchestra.run.mira083:Running: 'sudo logrotate /etc/logrotate.d/ceph-test.conf'
2016-12-07T18:00:44.611 INFO:teuthology.orchestra.run.mira100:Running: 'sudo logrotate /etc/logrotate.d/ceph-test.conf'
2016-12-07T18:00:44.614 INFO:teuthology.orchestra.run.mira105:Running: 'sudo logrotate /etc/logrotate.d/ceph-test.conf'
2016-12-07T18:00:44.887 INFO:teuthology.orchestra.run.mira105.stdout:skipping zero size read
2016-12-07T18:00:44.954 INFO:teuthology.orchestra.run.mira105.stdout:Size error: expected 0xa6f7c stat 0xd4000 seek 0xd5000
2016-12-07T18:00:44.954 INFO:teuthology.orchestra.run.mira105.stdout:LOG DUMP (2 total operations):
2016-12-07T18:00:44.954 INFO:teuthology.orchestra.run.mira105.stdout:1(1 mod 256): SKIPPED (no operation)
2016-12-07T18:00:44.954 INFO:teuthology.orchestra.run.mira105.stdout:2(2 mod 256): WRITE        0x1c748 thru 0xa6f7b    (0x8a834 bytes) HOLE
2016-12-07T18:00:44.990 INFO:teuthology.orchestra.run.mira105.stdout:Correct content saved for comparison
2016-12-07T18:00:44.990 INFO:teuthology.orchestra.run.mira105.stdout:(maybe hexdump "/home/ubuntu/cephtest/gmnt/test" vs "/home/ubuntu/cephtest/gmnt/test.fsxgood")
2016-12-07T18:00:45.000 INFO:teuthology.orchestra.run.mira105.stdout:
2016-12-07T18:00:45.000 INFO:teuthology.orchestra.run.mira105.stdout:===================================================================================
2016-12-07T18:00:45.000 INFO:teuthology.orchestra.run.mira105.stdout:=   BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES
2016-12-07T18:00:45.000 INFO:teuthology.orchestra.run.mira105.stdout:=   EXIT CODE: 120
2016-12-07T18:00:45.000 INFO:teuthology.orchestra.run.mira105.stdout:=   CLEANING UP REMAINING PROCESSES
2016-12-07T18:00:45.000 INFO:teuthology.orchestra.run.mira105.stdout:=   YOU CAN IGNORE THE BELOW CLEANUP MESSAGES
2016-12-07T18:00:45.000 INFO:teuthology.orchestra.run.mira105.stdout:===================================================================================
2016-12-07T18:00:45.000 INFO:teuthology.orchestra.run.mira105.stderr:[proxy:0:0@mira105] HYD_pmcd_pmip_control_cmd_cb (./pm/pmiserv/pmip_cb.c:886): assert (!closed) failed
2016-12-07T18:00:45.000 INFO:teuthology.orchestra.run.mira105.stderr:[proxy:0:0@mira105] HYDT_dmxu_poll_wait_for_event (./tools/demux/demux_poll.c:77): callback returned error status
2016-12-07T18:00:45.001 INFO:teuthology.orchestra.run.mira105.stderr:[proxy:0:0@mira105] main (./pm/pmiserv/pmip.c:206): demux engine error waiting for event
2016-12-07T18:00:45.002 INFO:teuthology.orchestra.run.mira105.stderr:[mpiexec@mira105] HYDT_bscu_wait_for_completion (./tools/bootstrap/utils/bscu_wait.c:76): one of the processes terminated badly; aborting
2016-12-07T18:00:45.002 INFO:teuthology.orchestra.run.mira105.stderr:[mpiexec@mira105] HYDT_bsci_wait_for_completion (./tools/bootstrap/src/bsci_wait.c:23): launcher returned error waiting for completion
2016-12-07T18:00:45.002 INFO:teuthology.orchestra.run.mira105.stderr:[mpiexec@mira105] HYD_pmci_wait_for_completion (./pm/pmiserv/pmiserv_pmci.c:217): launcher returned error waiting for completion
2016-12-07T18:00:45.002 INFO:teuthology.orchestra.run.mira105.stderr:[mpiexec@mira105] main (./ui/mpich/mpiexec.c:331): process manager error waiting for completion

I am not sure what the cause is. I'm leaving the test disabled for now and merging this PR.

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
4 years agofsx-mpi: remove chef task
Patrick Donnelly [Tue, 6 Dec 2016 19:41:00 +0000 (14:41 -0500)]
fsx-mpi: remove chef task

Per Sage's comment: https://github.com/ceph/ceph-qa-suite/pull/284#issuecomment-141463034

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
4 years agoMerge pull request #977 from ceph/wip-rgw-compression
Yehuda Sadeh [Wed, 7 Dec 2016 17:18:42 +0000 (09:18 -0800)]
Merge pull request #977 from ceph/wip-rgw-compression

rgw/verify: use rgw_compression_type = "random"

Reviewed-by: Yehuda Sadeh <yehuda@redhat.com>
Reviewed-by: Orit Wasserman <owasserm@redhat.com>
4 years agoupgrade/hammer-jewel-x/stress-split: no xenial
Sage Weil [Wed, 7 Dec 2016 15:27:19 +0000 (10:27 -0500)]
upgrade/hammer-jewel-x/stress-split: no xenial

Signed-off-by: Sage Weil <sage@redhat.com>
4 years agoMerge pull request #1211 from ceph/wip-15753-master
Sage Weil [Wed, 7 Dec 2016 14:54:24 +0000 (09:54 -0500)]
Merge pull request #1211 from ceph/wip-15753-master

upgrade/hammer-jewel-x suite

4 years agoupgrade/hammer-jewel-x: redo stress-split 1211/head
Sage Weil [Tue, 6 Dec 2016 22:46:10 +0000 (17:46 -0500)]
upgrade/hammer-jewel-x: redo stress-split

Make this easy: write a singel yaml that does the hammer install,
some limited work, then upgardes to jewel.  Copy it from the
parallel suite.  Then, symlink all of the rest from the jewel-x
stress-split suite.

Signed-off-by: Sage Weil <sage@redhat.com>
4 years agoupgrade/hammer-jewel-x: misc fixes
Sage Weil [Thu, 1 Dec 2016 23:09:11 +0000 (18:09 -0500)]
upgrade/hammer-jewel-x: misc fixes

4 years agoupgrade/hammer-jewel-x: fix upgrade order
Sage Weil [Thu, 1 Dec 2016 23:06:05 +0000 (18:06 -0500)]
upgrade/hammer-jewel-x: fix upgrade order

4 years agoupgrade/hammer-jewel-x: no xenial
Sage Weil [Thu, 1 Dec 2016 21:10:50 +0000 (16:10 -0500)]
upgrade/hammer-jewel-x: no xenial

Signed-off-by: Sage Weil <sage@redhat.com>
4 years agoupgrade/hammer-jewel-x: no mds, no fs tests
Sage Weil [Thu, 1 Dec 2016 20:48:13 +0000 (15:48 -0500)]
upgrade/hammer-jewel-x: no mds, no fs tests

4 years agoexclude ceph-mgr
Sage Weil [Thu, 1 Dec 2016 16:12:03 +0000 (11:12 -0500)]
exclude ceph-mgr

4 years agoupgrade/hammer-jewel-x/parallel: update
Sage Weil [Wed, 30 Nov 2016 20:00:42 +0000 (15:00 -0500)]
upgrade/hammer-jewel-x/parallel: update

Merger hammer-x (jewel branch) and jewel-x (master branch).

Signed-off-by: Sage Weil <sage@redhat.com>
4 years agoAdded new hammer-jewel-x suite
Yuri Weinstein [Mon, 17 Oct 2016 20:28:54 +0000 (20:28 +0000)]
Added new hammer-jewel-x suite

Removed rbd/test_librbd_python.sh for now
Removed all `sudo` from exec lines

Signed-off-by: Yuri Weinstein <yweinste@redhat.com>
4 years agotasks/cephfs: use kill in test_reconnect_eviction
John Spray [Wed, 7 Dec 2016 13:39:20 +0000 (13:39 +0000)]
tasks/cephfs: use kill in test_reconnect_eviction

Previously relied on client being able to unmount
while the MDS was offline, which is not necessarily
so.  Use kill instead.

Signed-off-by: John Spray <john.spray@redhat.com>
4 years agoMerge pull request #1300 from batrick/multimds-merge
John Spray [Wed, 7 Dec 2016 12:19:54 +0000 (12:19 +0000)]
Merge pull request #1300 from batrick/multimds-merge

suites/multimds: sync with fs suites

Reviewed-by: John Spray <john.spray@redhat.com>
4 years agosuites/multimds: add kclient disable override 1300/head
Patrick Donnelly [Fri, 2 Dec 2016 00:27:27 +0000 (19:27 -0500)]
suites/multimds: add kclient disable override

The libcephfs tests are negatively affected by other mounts. This commit
adds a kclient disable in addition to the ceph-fuse one.

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
4 years agosuites/multimds: sync with fs suites
Patrick Donnelly [Mon, 8 Aug 2016 02:25:19 +0000 (22:25 -0400)]
suites/multimds: sync with fs suites

This commit synchronizes the multimds suite with the fs suite. The
basic/verify sub-suites now do the same tests except with different
cluster layouts (i.e. multiple actives). This is mostly accomplished by
symlinking parts of each sub-suite to its counterpart in the fs suite.

This commit also does a few things of note to the prior multimds suite:

o Turn on directory fragmentation.
o Add several tests from fs/basic/tasks to multimds/basic.
o Remove libcephfs as fs/basic/tasks already contain
  multimds/basic/tasks.

Prior implementation and discussion are in PR#1114: https://github.com/ceph/ceph-qa-suite/pull/1114

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
4 years agosuites/kcephfs: disable broken python tests
John Spray [Wed, 7 Dec 2016 01:03:41 +0000 (01:03 +0000)]
suites/kcephfs: disable broken python tests

Disabling strays because of:
http://tracker.ceph.com/issues/17193

Disabling test_failover because of:
http://tracker.ceph.com/issues/18161

Signed-off-by: John Spray <john.spray@redhat.com>
4 years agotasks/cephfs: make test_journal_repair work with more daemons
John Spray [Mon, 14 Nov 2016 16:57:08 +0000 (16:57 +0000)]
tasks/cephfs: make test_journal_repair work with more daemons

Previously this assumed it was running with exactly two MDS
daemons.  When there were more, it would fail to execute
"fs reset" because the extra daemons were active in
the map.

Signed-off-by: John Spray <john.spray@redhat.com>
4 years agosuites/kcephfs: run python tests with 4 MDSs
John Spray [Tue, 8 Nov 2016 14:18:53 +0000 (14:18 +0000)]
suites/kcephfs: run python tests with 4 MDSs

For the benefit of test_failover

Signed-off-by: John Spray <john.spray@redhat.com>
4 years agotasks/cephfs: skip volumeclient client test on kcephfs
John Spray [Mon, 7 Nov 2016 18:15:01 +0000 (18:15 +0000)]
tasks/cephfs: skip volumeclient client test on kcephfs

This relies on quota-ish stuff that doesn't exist
in kclient.  We can still run the outer part
of the test though.

Signed-off-by: John Spray <john.spray@redhat.com>
4 years agotasks/cephfs: fix unmount call in volumeclient test
John Spray [Mon, 7 Nov 2016 13:58:53 +0000 (13:58 +0000)]
tasks/cephfs: fix unmount call in volumeclient test

Should have been to umount_wait, not umount (i.e.
the blocking foreground version).  This happened
to matter because umount_wait is more tolerant
of being called if the client is not already mounted.

Signed-off-by: John Spray <john.spray@redhat.com>
4 years agotasks/cephfs: special case for kcephfs behaviour
John Spray [Mon, 7 Nov 2016 12:00:25 +0000 (12:00 +0000)]
tasks/cephfs: special case for kcephfs behaviour

It does not generate EIOs in all the cases that
the fuse client does.

Signed-off-by: John Spray <john.spray@redhat.com>
4 years agotasks/cephfs: skip TestVolumeClient.test_evict on kcephfs
John Spray [Mon, 7 Nov 2016 11:57:49 +0000 (11:57 +0000)]
tasks/cephfs: skip TestVolumeClient.test_evict on kcephfs

This test was probably buggy and only happened to work
with ceph-fuse, because it expects the MDS to immediately
respond to updates to the client's auth caps, but that
doesn't happen.

Signed-off-by: John Spray <john.spray@redhat.com>
4 years agosuites/kcephfs: disable test_full for kcephfs
John Spray [Mon, 7 Nov 2016 10:11:12 +0000 (10:11 +0000)]
suites/kcephfs: disable test_full for kcephfs

Until http://tracker.ceph.com/issues/17204 is done.

Signed-off-by: John Spray <john.spray@redhat.com>
4 years agotasks/cephfs: fix race in TestStrays.test_dirfrag_limit
John Spray [Sun, 6 Nov 2016 22:58:33 +0000 (22:58 +0000)]
tasks/cephfs: fix race in TestStrays.test_dirfrag_limit

With the kernel client, this was proceeding to attempt
a split before the directory had persisted, because
there was no fsync after creating it.

Signed-off-by: John Spray <john.spray@redhat.com>
4 years agosuites/kcephfs: remove task for cap flush test
John Spray [Sun, 6 Nov 2016 22:24:16 +0000 (22:24 +0000)]
suites/kcephfs: remove task for cap flush test

This test skips all if not running against
fuse.

Signed-off-by: John Spray <john.spray@redhat.com>
4 years agotasks/cephfs: skip test_client_abort on kclient
John Spray [Sun, 6 Nov 2016 22:04:20 +0000 (22:04 +0000)]
tasks/cephfs: skip test_client_abort on kclient

This tests a fuse-only feature, ticket for adding
it in kclient is:
http://tracker.ceph.com/issues/17805

Signed-off-by: John Spray <john.spray@redhat.com>
4 years agotasks/cephfs: support multi-fs in kernel client
John Spray [Sun, 6 Nov 2016 22:01:00 +0000 (22:01 +0000)]
tasks/cephfs: support multi-fs in kernel client

Change the Mount interface to take it as an
argument to mount() instead of setting it
out of band in a config file as we used to
for the fuse client.

Signed-off-by: John Spray <john.spray@redhat.com>
4 years agotasks/cephfs: fix eviction test for volumeclient
John Spray [Sun, 6 Nov 2016 21:04:14 +0000 (21:04 +0000)]
tasks/cephfs: fix eviction test for volumeclient

It was not correct to expect a client to block after
eviction unless it was also deauthorised.  I guess
this was working with fuse because fuse does a less
good job at re-establishing a session than the kclient?

Signed-off-by: John Spray <john.spray@redhat.com>
4 years agotasks/cephfs: update test_journal_repair for kclient
John Spray [Sun, 6 Nov 2016 20:18:24 +0000 (20:18 +0000)]
tasks/cephfs: update test_journal_repair for kclient

Signed-off-by: John Spray <john.spray@redhat.com>
4 years agotasks/ceph: handle down nodes in log rotation
John Spray [Mon, 5 Sep 2016 10:21:24 +0000 (11:21 +0100)]
tasks/ceph: handle down nodes in log rotation

Signed-off-by: John Spray <john.spray@redhat.com>
4 years agotasks/cephfs: skip ENOSPC testing on kclient
John Spray [Fri, 2 Sep 2016 13:19:07 +0000 (14:19 +0100)]
tasks/cephfs: skip ENOSPC testing on kclient

Ticket #17204 exists to revive the old patchset, until then
we skip the tests.

Signed-off-by: John Spray <john.spray@redhat.com>
4 years agosuites/kcephfs: put clients on dedicated nodes
John Spray [Fri, 2 Sep 2016 12:58:25 +0000 (13:58 +0100)]
suites/kcephfs: put clients on dedicated nodes

Some tests want to use power cycling to reset stuck
mounts, and that needs to not kill OSDs as collateral
damage.

Need to revisit this to avoid unnecessarily using a whole
node for the client for those tests that don't require it
(i.e. those that don't use CephFSTestCase.REQUIRE_KCLIENT_REMOTE)

Signed-off-by: John Spray <john.spray@redhat.com>
4 years agotasks/cephfs: add a skip for non-fuse
John Spray [Fri, 2 Sep 2016 12:56:23 +0000 (13:56 +0100)]
tasks/cephfs: add a skip for non-fuse

Signed-off-by: John Spray <john.spray@redhat.com>
4 years agosuites/fs: s/yam/yaml/
John Spray [Thu, 1 Sep 2016 18:14:25 +0000 (19:14 +0100)]
suites/fs: s/yam/yaml/

Oops!

Signed-off-by: John Spray <john.spray@redhat.com>
4 years agosuites/kcephfs: s/yam/yaml/
John Spray [Thu, 1 Sep 2016 18:12:25 +0000 (19:12 +0100)]
suites/kcephfs: s/yam/yaml/

Oops!

Signed-off-by: John Spray <john.spray@redhat.com>
4 years agosuites/kcephfs: tolerate skipping tests
John Spray [Thu, 1 Sep 2016 18:11:35 +0000 (19:11 +0100)]
suites/kcephfs: tolerate skipping tests

With fuse client we expect everything to be
runnable so treat SkipTest as a failure because
it indicates something bad in the config.  With kclient
we known we have to skip some tests, so tolerate it.

Signed-off-by: John Spray <john.spray@redhat.com>
4 years agosuites/kcephfs: add TestFailover
John Spray [Thu, 1 Sep 2016 18:02:48 +0000 (19:02 +0100)]
suites/kcephfs: add TestFailover

Signed-off-by: John Spray <john.spray@redhat.com>
4 years agotasks/kclient: unmount conditionally
John Spray [Thu, 1 Sep 2016 17:40:21 +0000 (18:40 +0100)]
tasks/kclient: unmount conditionally

Don't assume the workload left the mount up and
running.

Signed-off-by: John Spray <john.spray@redhat.com>
4 years agotasks/cephfs: skip test_session_reject cleanly
John Spray [Thu, 1 Sep 2016 17:33:50 +0000 (18:33 +0100)]
tasks/cephfs: skip test_session_reject cleanly

Instead of asserting in configure_auth (which in fact
works fine with KernelMount.write_secret_file), raise
a SkipTest in test_session_reject (because the kernel
client cannot handle the client_metadata setting to
inject bogus data)

Signed-off-by: John Spray <john.spray@redhat.com>
4 years agotasks/cephfs: make KernelMount.umount_wait idempotent
John Spray [Tue, 30 Aug 2016 12:02:28 +0000 (13:02 +0100)]
tasks/cephfs: make KernelMount.umount_wait idempotent

Some places (notably CephFSTestCase.setUp) don't check
for mountedness before calling this.

Signed-off-by: John Spray <john.spray@redhat.com>
4 years agosuites/kcephfs: add python tests
John Spray [Wed, 10 Aug 2016 16:03:36 +0000 (17:03 +0100)]
suites/kcephfs: add python tests

Let's see which of these is happy running with a kernel client!

Signed-off-by: John Spray <john.spray@redhat.com>
4 years agoMerge pull request #1282 from batrick/fs-reorg
John Spray [Wed, 7 Dec 2016 00:44:31 +0000 (00:44 +0000)]
Merge pull request #1282 from batrick/fs-reorg

fs: unify common parts of sub-suites

Reviewed-by: Greg Farnum <gfarnum@redhat.com>
4 years agoMerge pull request #1275 from jcsp/wip-17853
John Spray [Wed, 7 Dec 2016 00:17:16 +0000 (00:17 +0000)]
Merge pull request #1275 from jcsp/wip-17853

Directory fragmentation tests

Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
4 years agotasks/cephfs: add test_standby_for_invalid_fscid
Ramakrishnan [Wed, 9 Nov 2016 09:31:55 +0000 (15:01 +0530)]
tasks/cephfs: add test_standby_for_invalid_fscid

Signed-off-by: Ramakrishnan Periyasamy <rperiyas@redhat.com>
4 years agoMerge pull request #532 from ceph/wip-12378
Sage Weil [Tue, 6 Dec 2016 18:38:44 +0000 (13:38 -0500)]
Merge pull request #532 from ceph/wip-12378

Adds task to wrap the aver utility

4 years agoMerge pull request #734 from dachary/wip-rados-thrash
Sage Weil [Tue, 6 Dec 2016 18:38:07 +0000 (13:38 -0500)]
Merge pull request #734 from dachary/wip-rados-thrash

openstack: rados/thrash...cache-pool-snaps.yaml: need 15GB

4 years agoMerge pull request #1019 from ceph/wip-rgw-fix-longpath
Sage Weil [Tue, 6 Dec 2016 18:37:11 +0000 (13:37 -0500)]
Merge pull request #1019 from ceph/wip-rgw-fix-longpath

rgw: split osd's in 2 nodes

4 years agoMerge pull request #980 from ceph/wip-fix-infinit
Sage Weil [Tue, 6 Dec 2016 18:05:27 +0000 (13:05 -0500)]
Merge pull request #980 from ceph/wip-fix-infinit

Fix infinite wait during monitor quorum check

4 years agorgw: configure compression in default zone when not multisite 977/head
Casey Bodley [Fri, 2 Dec 2016 19:48:44 +0000 (14:48 -0500)]
rgw: configure compression in default zone when not multisite

Signed-off-by: Casey Bodley <cbodley@redhat.com>
4 years agorgw/verify: set compression type=random
Casey Bodley [Thu, 28 Apr 2016 17:59:30 +0000 (13:59 -0400)]
rgw/verify: set compression type=random

instead of adding each compression type to the matrix for the rgw/verify
suite, use the new 'random' compression type to get test coverage of
the various compression plugins

this adds an 'rgw compression type' option to the client config, and later
adds that to the zone placement config

Signed-off-by: Casey Bodley <cbodley@redhat.com>
4 years agoFix infinite wait during monitor quorum check 980/head
Vasu Kulkarni [Mon, 2 May 2016 19:43:52 +0000 (12:43 -0700)]
Fix infinite wait during monitor quorum check

Signed-off-by: Vasu Kulkarni <vasu@redhat.com>
4 years agoSplit osd's into 2 nodes 1019/head
Vasu Kulkarni [Thu, 26 May 2016 17:44:07 +0000 (10:44 -0700)]
Split osd's into 2 nodes

http://tracker.ceph.com/issues/15612
Not all test machines(mira, vps) have 4 osd's and without this
change it will fail on nodes with < 4 osds, since one of the
osd will be on ext4 and it will hit long filename issue

Signed-off-by: Vasu Kulkarni <vasu@redhat.com>
4 years agoMerge pull request #535 from ceph/wip-12376
Sage Weil [Mon, 5 Dec 2016 22:53:51 +0000 (17:53 -0500)]
Merge pull request #535 from ceph/wip-12376

tasks/radosbenchsweep task

4 years agoMerge pull request #1024 from ceph/wip-default-fs
Sage Weil [Mon, 5 Dec 2016 22:53:01 +0000 (17:53 -0500)]
Merge pull request #1024 from ceph/wip-default-fs

tasks/ceph: default to xfs instead of root dir

4 years agoMerge pull request #1087 from sponce/master
Sage Weil [Mon, 5 Dec 2016 22:49:36 +0000 (17:49 -0500)]
Merge pull request #1087 from sponce/master

rados: Added tests for the striper layer of rados

4 years agoMerge pull request #956 from ceph/wip-rm-ext4
Samuel Just [Mon, 5 Dec 2016 22:42:08 +0000 (14:42 -0800)]
Merge pull request #956 from ceph/wip-rm-ext4

Remove ext4 option from rados:thrash tests

Reviewed-by: Samuel Just <sjust@redhat.com>
4 years agoMerge pull request #1301 from ceph/wip-valgrind-centos
Josh Durgin [Mon, 5 Dec 2016 19:36:45 +0000 (11:36 -0800)]
Merge pull request #1301 from ceph/wip-valgrind-centos

do all valgrind runs on centos

Reviewed-by: Josh Durgin <jdurgin@redhat.com>
4 years agodo all valgrind runs on centos 1301/head
Sage Weil [Mon, 5 Dec 2016 19:21:31 +0000 (14:21 -0500)]
do all valgrind runs on centos

The xenial valgrind has a bug with generating random
numbers; see http://tracker.ceph.com/issues/18126

Revert this when it is fixed.

Signed-off-by: Sage Weil <sage@redhat.com>
4 years agoMerge pull request #1299 from athanatos/wip-enable-hacky-flag
Sage Weil [Mon, 5 Dec 2016 16:45:13 +0000 (11:45 -0500)]
Merge pull request #1299 from athanatos/wip-enable-hacky-flag

ec-data-pool: enable the hacky_overwrites flag

4 years agoec-data-pool: enable the hacky_overwrites flag 1299/head
Samuel Just [Mon, 28 Nov 2016 23:25:05 +0000 (15:25 -0800)]
ec-data-pool: enable the hacky_overwrites flag

Signed-off-by: Samuel Just <sjust@redhat.com>
4 years agoMerge pull request #1298 from ceph/wip-pre-single-major-branch
Jason Dillaman [Fri, 2 Dec 2016 19:55:21 +0000 (14:55 -0500)]
Merge pull request #1298 from ceph/wip-pre-single-major-branch

krbd/unmap: pre-single-major is now 3.12.z

Reviewed-by: Jason Dillaman <dillaman@redhat.com>
4 years agorados/thrash/workloads: shorten rados bench a bit
Sage Weil [Fri, 2 Dec 2016 14:56:13 +0000 (09:56 -0500)]
rados/thrash/workloads: shorten rados bench a bit

I hit ENOSPC during thrashing on one run.

Signed-off-by: Sage Weil <sage@redhat.com>
4 years agoMerge pull request #1295 from dachary/wip-16506-ceph-coverage
Loic Dachary [Fri, 2 Dec 2016 10:42:03 +0000 (11:42 +0100)]
Merge pull request #1295 from dachary/wip-16506-ceph-coverage

upgrade: ceph-test is needed for ceph-coverage

Reviewed-by: Abhishek Lekshmanan <abhishek@suse.com>
4 years agoupgrade: ceph-test is needed for ceph-coverage 1295/head
Loic Dachary [Fri, 2 Dec 2016 09:27:25 +0000 (10:27 +0100)]
upgrade: ceph-test is needed for ceph-coverage

Do not exclude the ceph-test package otherwise the ceph-coverage
executable is not installed.

Fixes: http://tracker.ceph.com/issues/16506

Signed-off-by: Loic Dachary <loic@dachary.org>
4 years agoMerge pull request #1294 from ceph/wip-17960-master
Sage Weil [Thu, 1 Dec 2016 22:21:37 +0000 (17:21 -0500)]
Merge pull request #1294 from ceph/wip-17960-master

Removed xenial + Excluded librgw2

4 years agoupgrade/jewel-x/point-to-point: avoid xenial 1294/head
Sage Weil [Thu, 1 Dec 2016 22:21:19 +0000 (17:21 -0500)]
upgrade/jewel-x/point-to-point: avoid xenial

Xenial has newer packages in the distro and we want to
avoid fragile install instructions to avoid them.

Signed-off-by: Sage Weil <sage@redhat.com>
4 years agoExcluded librgw2
Yuri Weinstein [Thu, 1 Dec 2016 17:29:56 +0000 (17:29 +0000)]
Excluded librgw2
Fixes http://tracker.ceph.com/issues/17960

Signed-off-by: Yuri Weinstein <yweinste@redhat.com>
4 years agokrbd/unmap: pre-single-major is now 3.12.z 1298/head
Ilya Dryomov [Wed, 30 Nov 2016 11:16:58 +0000 (12:16 +0100)]
krbd/unmap: pre-single-major is now 3.12.z

3.13 won't build on newer distros, due to lack of gcc5 support.
nightly/pre-single-major (with a slash) is now gone.

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
4 years agoMerge pull request #1293 from ceph/wip-17977-fix
Dan Mick [Wed, 30 Nov 2016 23:15:15 +0000 (15:15 -0800)]
Merge pull request #1293 from ceph/wip-17977-fix

qemu: Set permissions on /dev/kvm

Reviewed-by: Dan Mick <dmick@redhat.com>
4 years agoqemu: Set permissions on /dev/kvm 1293/head
David Galloway [Tue, 22 Nov 2016 16:23:23 +0000 (11:23 -0500)]
qemu: Set permissions on /dev/kvm

Hack until fix in https://bugzilla.redhat.com/show_bug.cgi?id=1333159
makes it into CentOS repos.

Fixes: http://tracker.ceph.com/issues/17977

Signed-off-by: David Galloway <dgallowa@redhat.com>
4 years agoMerge pull request #1291 from ceph/wip-18085
Samuel Just [Wed, 30 Nov 2016 17:20:17 +0000 (09:20 -0800)]
Merge pull request #1291 from ceph/wip-18085

rados/thrash*: do no set osd_max_backfills=1

Reviewed-by: Samuel Just <sjust@redhat.com>
4 years agorados/thrash*: vary osd_max_backfills, always >1 wip-18085 1291/head
Sage Weil [Wed, 30 Nov 2016 17:13:14 +0000 (12:13 -0500)]
rados/thrash*: vary osd_max_backfills, always >1

This can lead to a copy-from vs backfill deadlock; see
http://tracker.ceph.com/issues/18085

This effectively reverts 5e880228fd7f59063d22a51eb5488b369b0c8360
which aimed to catch bugs in recovery reservations.

Signed-off-by: Sage Weil <sage@redhat.com>
4 years agoMerge pull request #894 from ceph/wip-bluestore
Sage Weil [Wed, 30 Nov 2016 02:55:46 +0000 (21:55 -0500)]
Merge pull request #894 from ceph/wip-bluestore

rados: test bluestore too

4 years agorados/objectstore/objectstore.yaml: do not test kstore 894/head
Sage Weil [Mon, 28 Nov 2016 21:10:22 +0000 (16:10 -0500)]
rados/objectstore/objectstore.yaml: do not test kstore

It's flaky.

Signed-off-by: Sage Weil <sage@redhat.com>
4 years agorados/objectstore/objectstore.yaml: test bluestore
Sage Weil [Mon, 21 Nov 2016 23:27:35 +0000 (18:27 -0500)]
rados/objectstore/objectstore.yaml: test bluestore

Signed-off-by: Sage Weil <sage@redhat.com>
4 years agorbd: test on filestore and bluestore
Sage Weil [Sat, 5 Nov 2016 02:51:14 +0000 (22:51 -0400)]
rbd: test on filestore and bluestore

Signed-off-by: Sage Weil <sage@redhat.com>
4 years agofs: test bluestore and filestore
Sage Weil [Sat, 22 Oct 2016 18:08:17 +0000 (14:08 -0400)]
fs: test bluestore and filestore

Signed-off-by: Sage Weil <sage@redhat.com>
4 years agorgw: test bluestore and filestore
Sage Weil [Sat, 22 Oct 2016 18:07:26 +0000 (14:07 -0400)]
rgw: test bluestore and filestore

Signed-off-by: Sage Weil <sage@redhat.com>
4 years agorados: test bluestore
Sage Weil [Wed, 21 Sep 2016 20:28:41 +0000 (16:28 -0400)]
rados: test bluestore

Signed-off-by: Sage Weil <sage@redhat.com>
4 years agoMerge pull request #1289 from ceph/wip-16397-workaround
John Spray [Wed, 30 Nov 2016 01:06:15 +0000 (01:06 +0000)]
Merge pull request #1289 from ceph/wip-16397-workaround

suites/knfs: pin to ubuntu

Reviewed-by: Jeff Layton <jlayton@redhat.com>
4 years agoMerge pull request #1283 from SUSE/wip-18014
Sage Weil [Tue, 29 Nov 2016 16:07:32 +0000 (11:07 -0500)]
Merge pull request #1283 from SUSE/wip-18014

thrashosds: try ceph-objectstore-tool for one minute

Reviewed-by: Loic Dachary <ldachary@redhat.com>
4 years agosuites/knfs: pin to ubuntu 1289/head
John Spray [Tue, 29 Nov 2016 14:28:13 +0000 (14:28 +0000)]
suites/knfs: pin to ubuntu

Workaround for http://tracker.ceph.com/issues/16397, which
manifests when we run bleeding edge kernels against the version
of nfs-utils in centos7.

Signed-off-by: John Spray <john.spray@redhat.com>
4 years agothrashosds: try ceph-objectstore-tool for 10 minutes 1283/head
Nathan Cutler [Thu, 24 Nov 2016 10:25:35 +0000 (11:25 +0100)]
thrashosds: try ceph-objectstore-tool for 10 minutes

If ceph-objectstore-tool binary is not present, it's likely because we're in
the middle of an upgrade. Do not try to run the binary until we verify that
it's really present. If it is absent, spend up to 10 minutes waiting for it to
appear.

Before this patch there was quite a large window for a race to occur. This
patch doesn't entirely eliminate it, but drastically reduces it.

Fixes: http://tracker.ceph.com/issues/18014
Signed-off-by: Nathan Cutler <ncutler@suse.com>