Commit Graph

4522 Commits

Author SHA1 Message Date
Sage Weil
bc9a66c7f4 upgrade/hammer-x/parallel: clients are hammer to for final workload
Signed-off-by: Sage Weil <sage@redhat.com>
2015-09-22 11:26:29 -04:00
Sage Weil
58453d3f55 upgrade/hammer-x/point-to-point: fix x client
- install hammer for hammer->x upgrade
- use hammer branch for workunit

Signed-off-by: Sage Weil <sage@redhat.com>
2015-09-22 11:26:29 -04:00
Sage Weil
1a0f68e0c9 upgrade/hammer-x/point-to-point: upgrade client for final test
Signed-off-by: Sage Weil <sage@redhat.com>
2015-09-21 08:17:08 -04:00
Sage Weil
eebd63a7e5 upgrade/hammer-x/point-to-point: use proper hammer test.sh
Signed-off-by: Sage Weil <sage@redhat.com>
2015-09-21 08:16:53 -04:00
Sage Weil
daf4f520c0 upgrade/hammer-x/parallel: upgrade client before final workload test
Signed-off-by: Sage Weil <sage@redhat.com>
2015-09-20 21:21:53 -04:00
Sage Weil
09318894a4 upgrade/hammer-x: fix full_sequential indentation
Broken by caa9235857

Signed-off-by: Sage Weil <sage@redhat.com>
2015-09-20 21:14:15 -04:00
Sage Weil
f6a96ad811 upgrade/hammer-x: use eastern tz for all upgrades
Hammer doesn't handle mixed timezones and hit_sets (cache pools)
correctly.  Avoid it.

Signed-off-by: Sage Weil <sage@redhat.com>
2015-09-18 17:31:06 -04:00
Sage Weil
a27cb8db3c timezone: add eastern, pacific, random timezone snippets
These need to be early tasks, probably. Random will pick either
eastern or pacific tz randomly.

Signed-off-by: Sage Weil <sage@redhat.com>
2015-09-18 17:29:31 -04:00
Sage Weil
eae42708f8 upgrade/hammer: do ubuntu for point to point
centos packages don't (git)build for these old
releases

Signed-off-by: Sage Weil <sage@redhat.com>
2015-09-18 11:26:30 -04:00
Sage Weil
3fd609f126 upgrade/hammer-x: only ubuntu
the centos packages don't (git)build

Signed-off-by: Sage Weil <sage@redhat.com>
2015-09-18 11:26:30 -04:00
Sage Weil
9993a2de84 Merge pull request #572 from dachary/wip-duplicate-centos7
do not duplicate CentOS 7 runs
2015-09-18 10:33:32 -04:00
Loic Dachary
99c6267473 do not duplicate CentOS 7 runs
The a-supported-distro.yaml file must be outside of the supported distro
directory otherwise it will have all suites including the distro
directory run CentOS 7 jobs twice.

Signed-off-by: Loic Dachary <loic@dachary.org>
2015-09-18 16:24:22 +02:00
Sage Weil
caa9235857 Merge pull request #566 from ceph/wip-s3-branch-fix
Use old branches as is without 'ceph-' for firefly, hammer etc
2015-09-18 10:10:10 -04:00
Sage Weil
a6fc7c6e1f upgrade/hammer-x: run rados.py to completion before moving on
Leave the final workloads alone.. those can run in parallel
with other stuff.

Signed-off-by: Sage Weil <sage@redhat.com>
2015-09-18 10:06:05 -04:00
Greg Farnum
4cccde634f Revert "Merge pull request #567 from ceph/ceph_fuse-timeout"
This reverts commit ff1655cb57, reversing
changes made to 2b25080d4f.

Since we haven't actually started the MDS daemons yet, this code is broken.

Signed-off-by: Greg Farnum <gfarnum@redhat.com>
2015-09-17 07:29:28 -07:00
Gregory Farnum
2f43e23fc6 Merge pull request #569 from dachary/wip-13031-packages
buildpackages: fix flake / pep8
2015-09-16 14:29:19 -07:00
John Spray
ff1655cb57 Merge pull request #567 from ceph/ceph_fuse-timeout
ceph: wait for CephFS to be healthy before proceeding

Reviewed-by: John Spray <john.spray@redhat.com>
2015-09-16 22:21:27 +01:00
Loic Dachary
e38a6d69df buildpackages: build the repository and the packages
http://tracker.ceph.com/issues/13031 Fixes: #13031

Signed-off-by: Loic Dachary <loic@dachary.org>
2015-09-16 23:19:34 +02:00
Greg Farnum
685d76a77c ceph: wait for CephFS to be healthy before proceeding
Signed-off-by: Greg Farnum <gfarnum@redhat.com>
2015-09-16 13:53:36 -07:00
Vasu Kulkarni
08bae3b9d2 Use old branches as is without 'ceph-' for firefly, hammer etc
Signed-off-by: Vasu Kulkarni <vasu@redhat.com>
2015-09-16 13:21:51 -04:00
Loic Dachary
8c618e5645 buildpackages: re-use config from install task
So that the packages built match what the install task expects.

http://tracker.ceph.com/issues/13031 Fixes: #13031

Signed-off-by: Loic Dachary <loic@dachary.org>
2015-09-15 17:37:05 +02:00
Loic Dachary
d40d8ff501 buildpackages: pep8 compliance
http://tracker.ceph.com/issues/13031 Fixes: #13031

Signed-off-by: Loic Dachary <loic@dachary.org>
2015-09-15 16:16:34 +02:00
Loic Dachary
0eeb441bb0 buildpackages: build ceph packages and upload them
Import install from task.

http://tracker.ceph.com/issues/13031 Fixes: #13031

Signed-off-by: Loic Dachary <loic@dachary.org>
2015-09-15 16:13:09 +02:00
Loic Dachary
2b25080d4f buildpackages: build ceph packages and upload them
Build Ceph packages from source for the required revision, os_type,
os_version and architecture and upload them to the gitbuilder
repository.

http://tracker.ceph.com/issues/13031 Fixes: #13031

Signed-off-by: Loic Dachary <loic@dachary.org>
2015-09-15 12:01:05 +02:00
Loic Dachary
8499d38095 Merge pull request #560 from dachary/wip-13026-hammer-stop
hammer-x: test CEPH_FEATURE_HAMMER_0_94_4
2015-09-15 11:45:11 +02:00
Sage Weil
95d5e38ef5 Merge pull request #564 from ceph/wip-hammer-x
fixes for hammer-x
2015-09-14 20:16:44 -04:00
Sage Weil
39f9ce6095 upgrade/hammer-x/split-*: ignore osdmap encode errors
This is normal during upgrade.

Signed-off-by: Sage Weil <sage@redhat.com>
2015-09-14 20:14:13 -04:00
Sage Weil
2460ce3bea upgrade/hammer-x: hammer test.sh before and during upgrade
Signed-off-by: Sage Weil <sage@redhat.com>
2015-09-14 20:14:13 -04:00
Loic Dachary
5befd50e84 hammer-x: test CEPH_FEATURE_HAMMER_0_94_4
http://tracker.ceph.com/issues/13026 Refs: #13026

Signed-off-by: Loic Dachary <loic@dachary.org>
2015-09-13 00:46:03 +02:00
Sage Weil
4977d3945e Merge pull request #563 from ceph/wip-13060
fix test for 11429 (and now 13060)
2015-09-12 09:24:33 -04:00
Sage Weil
f0c925e30a suites/rados/singleton-nomsgr/all/11429.yaml: double-hop and fix
- simplify this.. lots of extra cruft we don't need
- restart twice at hammer to ensure that we can still load pgs
  post-upgrade
- do the same for the final version.

Fixes: #11429 (again, for ~infernalis)
Fixes: #13060
Signed-off-by: Sage Weil <sage@redhat.com>
2015-09-12 09:22:30 -04:00
Zack Cerza
37559f50dc Merge pull request #562 from ceph/wip-radosbench-cleanup
tasks/radosbench: allow cleanup: false
2015-09-11 14:55:05 -06:00
Sage Weil
5af28cf746 tasks/radosbench: allow cleanup: false
Signed-off-by: Sage Weil <sage@redhat.com>
2015-09-11 16:52:41 -04:00
Zack Cerza
e9847570de Merge pull request #561 from ceph/wip-sudo
sudo ceph
2015-09-11 10:20:50 -06:00
Zack Cerza
22dc3df4ab Merge pull request #559 from ceph/wip-10328
leaks
2015-09-11 10:19:46 -06:00
Sage Weil
5eb0e4dd2f suites: sudo ceph
Signed-off-by: Sage Weil <sage@redhat.com>
2015-09-11 12:15:01 -04:00
Sage Weil
dad981d339 tasks: sudo ceph for cli
/var/run/ceph is 770.  This is mainly necessary for any
interaction with the daemon sockets, but it is what users do
and it may avoid log noise.

Signed-off-by: Sage Weil <sage@redhat.com>
2015-09-11 12:15:01 -04:00
Sage Weil
822fa666dd rados/singleton-nomsgr/all/valgrind-leaks: verify leak checking works
Fixes: #10328
Signed-off-by: Sage Weil <sage@redhat.com>
2015-09-11 11:53:37 -04:00
Sage Weil
9b3f36f91f ceph: add option to expect valgind errors and fail if there are none
See #10328
Signed-off-by: Sage Weil <sage@redhat.com>
2015-09-11 11:53:37 -04:00
Loic Dachary
45c9c0aed0 distros: alias a-supported-distro.yaml to CentOS 7
Some jobs are distribution agnostic and it is meaningless to run them on
all supported distributions. For such jobs it is convient to require
than they run on a supported distribution but we don't care which. This
can be done by including a-supported-distro.yaml in the suite sub
directory instead of linking to distros/supported.

Signed-off-by: Loic Dachary <loic@dachary.org>
2015-09-11 12:12:00 +02:00
Sage Weil
88e4a2451a Merge pull request #558 from ceph/wip-12973
rados: 'failed to encode ...' warnings are normal on upgrades

Tested-by: Kefu Chai <kchai@redhat.com>
2015-09-10 08:57:59 -04:00
Sage Weil
3084453b04 rados: 'failed to encode ...' warnings are normal on upgrades
Fixes: #12973
Signed-off-by: Sage Weil <sage@redhat.com>
2015-09-09 22:04:13 -04:00
Sage Weil
edf36f971d Merge pull request #542 from ceph/wip-12551-cephfs-truncate-sec
Adds test for ceph/pull/5442

Reviewed-by: Sage Weil <sage@redhat.com>
2015-09-09 16:22:38 -04:00
Ivo Jimenez
d83a95ea3e suites/rados/singleton-nomsgr/all/cache-fs-trunc.yaml
Verify the cache tier doesn't clobber truncate_{seq,size}
2015-09-09 16:21:39 -04:00
Josh Durgin
3e5b091fc5 Merge pull request #557 from ceph/wip-12563
suites/hammer/older: do not upgrade client while running RBD import/export

Reviewed-by: Josh Durgin <jdurgin@redhat.com>
2015-09-09 11:44:11 -04:00
Gregory Farnum
6ac2774823 Merge pull request #550 from ceph/wip-12806
tasks/qemu: avoid adding duplicated entries to /etc/exports

Reviewed-by: Greg Farnum <gfarnum@redhat.com>
2015-09-09 14:53:45 +01:00
Jason Dillaman
cb7eb2ad6f suites/hammer/older: do not upgrade client while running RBD import/export
Between the Giant and Hammer releases, the symbols exported from librados/librbd
were cleaned up.  This creates a short window of time between installing
upgraded packages where symbol lookup errors could occur.

Fixes: #12563
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
2015-09-09 09:09:31 -04:00
Sage Weil
2019a78cb5 Merge pull request #556 from ceph/wip-stuck-pg-dump
tasks/ceph_manager: dump pgs when recover times out

Reviewed-by: Loic Dachary <ldachary@redhat.com>
2015-09-08 09:12:53 -04:00
Sage Weil
a328e3e60d tasks/ceph_manager: dump pgs when recover times out
It is really hard to map a stuck recovery back to the pgs that
are stuck.  This will make it easy.

Signed-off-by: Sage Weil <sage@redhat.com>
2015-09-08 08:59:49 -04:00
Loic Dachary
6eb73595cd Merge pull request #541 from XinzeChi/wip-rebug
fix rebug rados

Reviewed-by: Loic Dachary <ldachary@redhat.com>
2015-09-07 09:56:51 +02:00