vstart_runner can't find arguments to ceph daemons to identify them with
ps -x because commands are cut off at terminal width. Add -ww for wide
output.
Signed-off-by: Douglas Fuller <dfuller@redhat.com>
Fortunately we already have a test that creates the condition,
so just tweak it to exceed the 150% threshold for the health warning,
and check that the health message appears.
Signed-off-by: John Spray <john.spray@redhat.com>
Test the usecases for the authentication metadata stored
by the volume client:
* Obtain the list of auth IDs having access to a volume.
* Restrict volume access to auth IDs of a single (OpenStack)
tenant to enforce strong tenant isolation of volumes.
Signed-off-by: Ramana Raja <rraja@redhat.com>
Message is logged as, filesystem is mounted,
despite the vstart_runner just trying to mount
at this stage.
Signed-off-by: Ramana Raja <rraja@redhat.com>
... instead of iterating over all blacklist entries. Fall
back to the old way if the new way doesn't work (i.e. on
old ceph)
Signed-off-by: John Spray <john.spray@redhat.com>
``long_running`` needs a better name, it's really just a
marker that says a test is for use in teuthology but not vstart.
Signed-off-by: John Spray <john.spray@redhat.com>
The fragment configuration uses 10000 for the fragment max size. The reason for
this is that many tests add 1000 files to a single directory which will hit
this limit without fragmentation catching up.
The test_dirfrag_limit test confirms:
o That the directory fragment size cannot exceed mds_bal_fragment_size_max (using a limit of 50 in all configurations).
o That fragmentation (forced) will allow more entries to be created.
o That unlink fails when the stray directory fragment becomes too large and that unlinking may continue once those strays are purged.
Tests: https://github.com/ceph/ceph/pull/9789
Issue: http://tracker.ceph.com/issues/16164
Signed-off-by: Patrick Donnelly <batrick@batbytes.com>
Dropping the 'x' caused the ps lines not to include
the "-i <myid>" part of the command line that we use
to find named daemons' pids.
Signed-off-by: John Spray <john.spray@redhat.com>
Even though check_subs was broken, the multiclient
test was still passing because it didn't test that
clients survived a failover in the multi-filesystem
case.
Fixes: http://tracker.ceph.com/issues/16022
Signed-off-by: John Spray <john.spray@redhat.com>
Don't mistake the daemon-helper line for the actual
daemon process.
Fixes: http://tracker.ceph.com/issues/15900
Signed-off-by: John Spray <john.spray@redhat.com>
Test whether mount access to volumes can be restricted to read-only.
Fixes: http://tracker.ceph.com/issues/15614
Signed-off-by: Ramana Raja <rraja@redhat.com>
In the test_volume_client module, refactor code to add a function
that configures credentials for a guest to mount a volume. The
function could be used by tests to setup a guest client to mount
a volume with fewer lines of code.
Signed-off-by: Ramana Raja <rraja@redhat.com>
Now that daemon commandlines include "--cluster" they
easily overflow ps's default behaviour.
Fixes: http://tracker.ceph.com/issues/15857
Signed-off-by: John Spray <john.spray@redhat.com>
For reproducing error handling paths during purge,
such as #15266 (ASCII exception)
Fixes: http://tracker.ceph.com/issues/15266
Signed-off-by: John Spray <john.spray@redhat.com>
Also ensure namespace_prefix and vol_prefix are set to default values
if they are not explicitly passed.
Fixes#15417
Signed-off-by: Xiaoxi Chen <xiaoxchen@ebay.com>
Test whether the CephFSVolumeClient can evict a client based on its
auth ID and the volume path it has mounted.
Fixes: http://tracker.ceph.com/issues/15045
Signed-off-by: Ramana Raja <rraja@redhat.com>
The mount points of the test clients were created based on the auth
ID they used. So two clients using the same auth ID could not create
different mount points in a straight forward way. Expose the
directory name of the mount point as a public attribute allowing
client mount points to be easily changed whenever needed.
Signed-off-by: Ramana Raja <rraja@redhat.com>
For the case where we have a daemon come up that
wants to be a standby replay, but someone else
is already following the target, so he has to
be just a regular standy instead.
Signed-off-by: John Spray <john.spray@redhat.com>
Previously, an MDS could be used as a replay even
if it had "standby replay = false", this is changed
now so that replays will only happen explicitly.
Signed-off-by: John Spray <john.spray@redhat.com>
