A first step to do more automatic code checks on the qa/
directory. This is useful while transitioning to python3.
Also use log_exc to top-level to not run into:
error: Argument 1 to "log_exc" has incompatible type
"Callable[[OSDThrasher], Any]"; expected "OSDThrasher"
Signed-off-by: Thomas Bechtold <tbechtold@suse.com>
client,common,mgr,rbd: clang related cleanups
Reviewed-by: Willem Jan Withagen <wjw@digiware.nl>
Reviewed-by: Jason Dillaman <dillaman@redhat.com>
Reviewed-By: Neha Ojha <nojha@redhat.com>
Reviewed-By: Patrick Donelly <pdonell@redhat.com>
* refs/pull/33730/head:
cephadm/vstart-smoke.sh: a smoke test to run against vstart
mgr/cephadm: fix prom reconfig loop when dashboard is disabled
mgr/cephadm: clean up log messages a bit
mgr/cephadm: force remove daemons when removing service
mgr/cephadm: do not error out on service rm|action if no daemons
mgr/orchestrator: allow 'orch rm mon|mgr --force'
Reviewed-by: Michael Fritch <mfritch@suse.com>
The device inventory rarely changes.
A daemon inventory of 60s may trigger every time around the loop when
doing upgrades, which is pretty slow (esp since it's currently done
synchronously). We're doing a good job (I think?) of queueing refreshes,
so the only reason to time this data out is to catch daemon stops that
are caused by external factors.
Signed-off-by: Sage Weil <sage@redhat.com>
This provides decent coverage of cephadm in practice, and can run against
a vstart cluster without having to go build containers.
vstart-cleanup.sh will clean up an aborted test reasonably well without
requiring sudo.
Caveats:
- there is no coverage for osds at this point.
Signed-off-by: Sage Weil <sage@redhat.com>
We need to consider the mgrs a dep even if the prometheus module isn't
enabled, so that we are consistent with _calc_daemon_deps() and don't
reconfig prom in a loop.
Signed-off-by: Sage Weil <sage@redhat.com>
The rm CLI checks for --force based on type, so it isn't passed down
through the implementations. So, pass it through.
Signed-off-by: Sage Weil <sage@redhat.com>
* refs/pull/33704/head:
mgr/cephadm: avoid racing updates to spec dict
mgr/cephadm: make _do_upgrade work synchronously
mgr/cephadm/tests/test_cephadm: remove broken rgw_update_fail test
mgr/cephadm: make _apply_service move services
mgr/cephadm: log when we create, destroy daemons
mgr/cephadm: drop async_map_completion decorators from create funcs
mgr/cephadm: _create_osd -> _create_osds
mgr/cephadm: _remove_daemon -> _remove_daemons
Reviewed-by: Joshua Schmid <jschmid@suse.de>
Reviewed-by: Michael Fritch <mfritch@suse.com>
Remove Octopus from list of active stable releases (it's not stable yet)
and put the 15.1.0 link in the Development column.
Signed-off-by: Nathan Cutler <ncutler@suse.com>
This was trying to show that we won't put 2 RGWs from different zones
on the same host. In reality, the test was only passing because it was
specifying an invalid placement spec (count=3 in a 2-host cluster).
It's true that we shouldn't do that because they won't be able to bind,
but the HostAssignment code doesn't handle this case currently.
See https://tracker.ceph.com/issues/44397
Signed-off-by: Sage Weil <sage@redhat.com>
* refs/pull/33565/head:
os/bluestore: log allocation stats on a daily basis.
Reviewed-by: Adam Kupczyk <akucpzyk@redhat.com>
Reviewed-by: Sage Weil <sage@redhat.com>
We are sending a request that returns an error *and* a data payload. In order
to reliably get the payload, we need to set the RETURNVEC flag so that it is
recorded in the pg log. Otherwise, we might get unlucky and resend the request
and end up in the dup handler and get the error code but no payload.
Fixes: https://tracker.ceph.com/issues/44385
Signed-off-by: Sage Weil <sage@redhat.com>
mgr/dashboard: Not able to restrict bucket creation for new user
Reviewed-by: Tatjana Dehler <tdehler@suse.com>
Reviewed-by: Tiago Melo <tmelo@suse.com>
* refs/pull/33705/head:
qa/suites/upgrade/nautilus-x/parallel: restart mgr.x before mons
Reviewed-by: Sage Weil <sage@redhat.com>
Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
Reviewed-by: Josh Durgin <jdurgin@redhat.com>
