We were waiting for teh dameon to fail, not for the mons to notice,
which meant that after we restart the daemon the next tell could
try to talk to the old osd.
Signed-off-by: Sage Weil <sage@redhat.com>
The thrasher needs disk attached to run against xfs, ext4 or btrfs. And
some jobs use more disks and do not fit in 40GB.
http://tracker.ceph.com/issues/13450Fixes: #13450
Signed-off-by: Loic Dachary <loic@dachary.org>
If given only 8GB RAM, ceph_test_msgr may abort with buffer::bad_alloc.
http://tracker.ceph.com/issues/11260Fixes: #11260
Signed-off-by: Loic Dachary <loic@dachary.org>
A new test verifies that we are stopped by the pool quota (and get
the right error messages or block). See ceph.git
32962740ce.
Signed-off-by: Sage Weil <sage@redhat.com>
Blackhole filestore ops so that we ensure it doesn't complete
the pg deletions before the restart function does a clean shutdown
etc.
Signed-off-by: Sage Weil <sage@redhat.com>
Restart can be slow enough that osd.1 and 2 finish deleting the
pgs. Verifying one osd sees the instance is sufficient.
Signed-off-by: Sage Weil <sage@redhat.com>
Link the distro directory to the directory containing all supported
distros. Add the x86_64 arch contraint required by the isa plugin to an
isolated file that is combined with all jobs.
Signed-off-by: Loic Dachary <loic@dachary.org>
- 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>
- in general, test simple vs async vs random
- not for msgr-less workloads
- not for thrash-erasure-*.. the regular thrash
should cover it.
Signed-off-by: Sage Weil <sage@redhat.com>
Add a workload that uses the lrc erasure code plugin. Instead of adding
it to suites/rados/thrash-erasure-code/workloads, a new suite is created
at suites/rados/thrash-erasure-code-big because it needs more OSDs than
other erasure code plugins. The alternative would be to increase the
number of OSDs for all erasure code plugins, but that would needlessly
increase the resources requirements.
* cluster/12-osds.yaml creates a 12 OSDs, 3 MONs cluster
* thrash-erasure-code-big/thrashers/*.yaml are the same as
thrash-erasure-code/thrashers/*.yaml except they require that at
least 8 OSDs are in at all times (instead of 4) because lrc PGs with
k=4, m=2, l=3 are undersized if they do not have 8 OSDs. It is
possible that crush fails to map 8 OSDs when only 8 OSDs are
available, but that must not disturb the workload because min_size is
4.
http://tracker.ceph.com/issues/11666Fixes: #11666
Signed-off-by: Loic Dachary <ldachary@redhat.com>
