so we can drain all pending tasks referencing messenger before
destroying it, otherwise they will be using a already-stopped
messenger when they are still alive.
Signed-off-by: Kefu Chai <kchai@redhat.com>
* refs/pull/31828/head:
mgr/orchestrator: include container_image in 'service ls' output
ceph-daemon: include container_image_name and container_image_id in 'ls' output
common/options: default to ceph/ceph:latest-master-devel (for now)
Reviewed-by: Michael Fritch <mfritch@suse.com>
Reviewed-by: Sebastian Wagner <swagner@suse.com>
* refs/pull/31864/head:
osd/PeeringState: clear LAGGY and WAIT states on exiting Started
Reviewed-by: Neha Ojha <nojha@redhat.com>
Reviewed-by: Samuel Just <sjust@redhat.com>
Keep the host list in an explicit 'inventory' key, and provide a dict for
each host so that we can keep other metadata (like labels).
Signed-off-by: Sage Weil <sage@redhat.com>
* refs/pull/31567/head:
common: move gen_rand_alphanumeric() helpers into common
test: add unit tests for gen_rand_alphanumeric helpers
Reviewed-by: Adam C. Emerson <aemerson@redhat.com>
* refs/pull/31781/head:
cmake/modules: look for GNU make first instead of BSD make
cmake: support parallel build for rocksdb
Reviewed-by: Willem Jan Withagen <wjw@digiware.nl>
Reviewed-by: Kefu Chai <kchai@redhat.com>
* refs/pull/31775/head:
mds: remove the incorrect comments
Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
* refs/pull/31762/head:
mgr/ssh: improve logging for mds removal
mgr/volumes: move up 'confirm' validation
Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
Reviewed-by: Tim Serong <tserong@suse.com>
Reviewed-by: Jos Collin <jcollin@redhat.com>
These keys were a mirror of the mon's purged_epoch keys. We were writing
them, but never looking at them. We only need the per-pool keys that
are indexed by snapid for lookups, so we should stop storing the per-epoch
records.
Note that the OSDSuperblock has the purged_snaps_last that serves as our
position/cursor for ingesting the mon keys.
Fixes: https://tracker.ceph.com/issues/42012
Signed-off-by: Sage Weil <sage@redhat.com>
The removal of remote mirror image status records no longer
(accidentally) searches through nearly the full omap database and
getting the status of a single image no longer involves duplicate
lookups (to pull the fsid keys and then to read the data).
Fixes: https://tracker.ceph.com/issues/42576
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
These flags were not getting cleared except in recheck_readable(), which
meant that a flag from a prior interval could bleed into a new interval.
More dangerously, in a mixed-version cluster, one interval might include
all octopus+ OSDs while the next might include a pre-octopus OSD, bypassing
most of the laggy recheck code. This could lead to a stalled request
and/or requeue ordering bug when release_object_locks() looked at
is_laggy() and put a lock waiter on the waiting_for_readable list.
Fixes: https://tracker.ceph.com/issues/42978
Signed-off-by: Sage Weil <sage@redhat.com>
`seastar::future<T>::available()` already verifies whether
a future instance stores exception inside. It's unnecessary
to ensure that with further call to `failed()`.
This tiny clean-up removes the duplicative check in
`with_blocking_future()`.
Signed-off-by: Radoslaw Zarzynski <rzarzyns@redhat.com>
