* add mgr.x to roles
* to setup mgr and set the require-osd-release bit in osdmap
* do not restart an osd for waiting for healthy: the cluster is not
healthy until the require-osd-release=luminous is set in osdmap, and a
mgr is up and running.
Fixes: http://tracker.ceph.com/issues/20342
Signed-off-by: Kefu Chai <kchai@redhat.com>
Set pool size back to 2 so we don't have to have backfill
complete (despite rejection probability) in order to get back to
healthy. This way we scrub on cleanup.
Signed-off-by: Sage Weil <sage@redhat.com>
If we leave the quota set, the proxied ops will block
indefinitely, which will block scrubbing on the cache tier pgs
indefinitely.
Signed-off-by: Sage Weil <sage@redhat.com>
This reverts commit be88220965.
we can workaround the limit of old requests by using the backward
compatible way.
Signed-off-by: Kefu Chai <kchai@redhat.com>
These did not do enough metadata operations to
trigger any exports between MDS ranks, so it is
not useful to have them running in the multimds
suite.
Signed-off-by: John Spray <john.spray@redhat.com>
...so that we can't "pass" the test by accidentally
failing to ever export anything automatically
like in http://tracker.ceph.com/issues/20131
Signed-off-by: John Spray <john.spray@redhat.com>
This reverts 693bd23851, which was
added in response to http://tracker.ceph.com/issues/18126. But
we updated the Ubuntu packages in sepia so it should be good to go.
Signed-off-by: Greg Farnum <gfarnum@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 default tunables for luminous allow straw2 buckets, which aren't
supported by pre-single-major.yaml kernel. The map is generated before
crush_tunables override takes effect, so set the bucket algorithm to
straw manually.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
This parsed out as
tasks:
- install: null
- ceph:
conf:
osd: osd max object name len = 400 osd max object namespace len = 64
- workunit:
clients:
all:
- rados/test_health_warnings.sh
which is clearly not correct.
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>
