When we are doing cache tiering, we are more sensitive to short PG logs
because the dup op entries are not perfectly promoted from the base to
the cache.
See:
http://tracker.ceph.com/issues/38358http://tracker.ceph.com/issues/24320
This works around the problem by not testing short pg logs in combination
with cache tiering. This works because the short_pg_log.yaml fragment
sets the short log in the [global] section but the cache workloads overload
it (back to a large/default value) in the [osd] section.
Signed-off-by: Sage Weil <sage@redhat.com>
Seeing some hangs when the mon is forwarding mgr commands (pg deep-scrub)
to the mgr. This is a buggy test (it should send it to the mgr directly)
but it is helpful to verify the mon forwarding behavior works.
Signed-off-by: Sage Weil <sage@redhat.com>
With automatic balancing on, and if mode is set to upmap,
balancer will fail silently if min_compat_client is lower than
luminous.
You can't figure out that unless you take a closer look at the
mgr log, which is super annoying..
Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
This utilizes the recent feature in teuthology [1] to skip hidden files in
suites when building the job matrix.
Idea of this change is to enable referring to the top-level qa directory in a
position-independent way such that copies of a suite to another location do not
break any symlinks.
[1] https://github.com/ceph/teuthology/pull/1185
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
pg will be created when increasing pgp-num and pg-num. so at that
moment, PG_AVAILABILITY is reported. so whitelist it in all tests which
run rados/test.sh. that script exercises ceph_test_rados_api_list.
Fixes: http://tracker.ceph.com/issues/23763
Signed-off-by: Kefu Chai <kchai@redhat.com>
1.add tier_promote op for redirect and chunked cases.
2.rename set-chunk.yaml due to current chunked object
only for the read case.
Signed-off-by: Myoungwon Oh <omwmw@sk.com>
current chunked object and ChunkReadOp are
only for the read case.
write op and promote_object() still be tested without ChunkReadOp
by another ceph_test_rados in the same test suite (with --set_chunk)
Signed-off-by: Myoungwon Oh <omwmw@sk.com>
We can't mix the balancer compat-set testing with firefly tunables because
it requires that all buckets be straw2.
Signed-off-by: Sage Weil <sage@redhat.com>
- snapdir conversion (at-end) stuff
- merge luminous-specific collections that avoided the above back
into their normal locations
Signed-off-by: Sage Weil <sage@redhat.com>
so we can avoid the warnings like
grep: Unmatched ( or \(
because we pass the whitelisted string to `egrep -v "$1"` directly.
Signed-off-by: Kefu Chai <kchai@redhat.com>
With the peering deletes change, setting luminous sets the osdmap flag
which triggers a new peering interval. That can lead to health warnings
about PG_AVAILABILITY or PG_DEGRADED. Ignore those!
Fixes: http://tracker.ceph.com/issues/20693
Signed-off-by: Sage Weil <sage@redhat.com>
1) ruleset is an obsolete term, and
2) crush-{rule,failure-domain,...} is more descriptive.
Note that we are changing the names of the erasure code profile keys
from ruleset-* to crush-*. We will update this on upgrade when the
luminous flag is set, but that means that during mon upgrade you cannot
create EC pools that use these fields.
When the upgrade completes (users sets require_osd_release = luminous)
existing ec profiles are updated automatically.
Signed-off-by: Sage Weil <sage@redhat.com>
- OSDMap encode and decode translate between the flags and int
representations.
- OSDMap::Incremental only does decode; we do not expect to ever encode
an incremental osdmap for an old osd that sets any of these flags.
- the 'osd set' command still lets you set the jewel and kraken flags,
but not luminous.
- OSDMap::apply_incremental handles the conversion of legacy require flags
to the new field if the jewel or kraken flags have to be set before
starting the osd upgrade.
- clear out the legacy flags when we make the luminous transition only;
until then we keep using the old flag in the encoded and decoded version
(although the require_osd_release field will be accurate in memory in all
cases).
Signed-off-by: Sage Weil <sage@redhat.com>
The scrub_pgs command also waits for healthy for a while, but fails
silently if it times out, which means the subsequent scrubs will also
fail to clean up.
This forces an earlier failure that does not obscure the root cause.
Signed-off-by: Sage Weil <sage@redhat.com>
The OSDs must have a map reflecting the require_luminous flag in order
for the legacy conversion to happen. A quick rados bench should ensure
that.
Signed-off-by: Sage Weil <sage@redhat.com>
This lets us run multiple cleanup steps right before ceph
teardown.
Note that we drop the facet from multimon/ because it
doesn't factor out cluster creation before this step
properly. That's fine because the require_luminous
cleanup shouldn't be related to the multimon tests.
Signed-off-by: Sage Weil <sage@redhat.com>
