]> git.apps.os.sepia.ceph.com Git - ceph.git/commit
qa/cephfs: add session_timeout option support 37841/head
authorXiubo Li <xiubli@redhat.com>
Tue, 20 Oct 2020 05:26:33 +0000 (01:26 -0400)
committerNathan Cutler <ncutler@suse.com>
Tue, 27 Oct 2020 08:31:45 +0000 (09:31 +0100)
commitf66b52806a7107465886e9b4a6fae83f517fdca0
treeb58e2a78c81be14b064f8569a0cc6c7197db49be
parent5341b65ff314329a8b44ae78862001d6c1b46794
qa/cephfs: add session_timeout option support

When the mds revoking the Fwbl caps, the clients need to flush
the dirty data back to the OSDs, but the flush may make the OSDs
to be overloaded and slow, which may take more than 60 seconds to
finish. Then the MDS daemons will report the WRN messages.

For the teuthology test cases, let's just increase the timeout
value to make it work.

Fixes: https://tracker.ceph.com/issues/47565
Signed-off-by: Xiubo Li <xiubli@redhat.com>
(cherry picked from commit 0422673b6150df851a4ea1662637a77585cde52d)
qa/cephfs/overrides/session_timeout.yaml [new file with mode: 0644]
qa/suites/fs/basic_workload/overrides/session_timeout.yaml [new symlink]
qa/suites/fs/thrash/overrides/session_timeout.yaml [new symlink]
qa/suites/fs/verify/overrides/session_timeout.yaml [new symlink]
qa/tasks/ceph.py
qa/tasks/cephfs/filesystem.py
qa/tasks/vstart_runner.py