This fixes errors caused by remount done by some tests (test_recovery_pool.py)
where the fs name is not given.
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
The MDS may not be on the same machine where the cluster command is run.
Fixes: http://tracker.ceph.com/issues/24858
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
A change to the exact-match behavior in leaf stat operations is
required to support instantiation of a full path hierarchy when
one or more intervening directories has not been enumerated.
Fixes: http://tracker.ceph.com/issues/24915
Signed-off-by: Matt Benjamin <mbenjamin@redhat.com>
* refs/pull/21885/head:
qa: update cluster log health warning message
qa: add tests for client features
mds: evict clients that lack required features
mds: cleanup MDSRank::evict_client
mds: infer client version by client metadata and connection's features
mds: introduce "ceph fs set <fs_name> min_compat_client <release_name>"
mds: tell client why it's rejected
mds: introduce cephfs' own feature bits
mds: make Server::prepare_force_open_sessions() update client metadata
Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
The existing logic appears able to cause propagation of a failed
exclusive create to the client, when it should instead have been
retried, due to disagreement about the logical write offset. (The
value of ofs here could be > 0 due to the operation of a stacked
write filter [e.g., compressor], when the RADOS write offset was
0 and hence an exclusive write that should be retried).
Rationale for fix by Casey.
http://tracker.ceph.com/issues/22790
Signed-off-by: Matt Benjamin <mbenjamin@redhat.com>
a single call to time_log_trim() will trim a maximum of 1000 entries.
now loops until a return code of -ENODATA indicates that all entries
in the given range were trimmed
an optional --trim-delay-ms option can be used to limit the frequency
of osd ops
Fixes: http://tracker.ceph.com/issues/24873
Signed-off-by: Casey Bodley <cbodley@redhat.com>
we can just filter out ceph options from command line args instead of
parsing config in env variables, conf files and argvs.
the config-parsing steps were modeled after the ones in rbd-nbd.cc, but
the whole purpose of reading configurations from env/conf-file is to get
the "rbd_default_pool" for setting cfg->poolname. apparently, this does
not apply to rbd-ggate, where global_init() will take care of all the
trivial work to populate the settings. moreover, `config` is not
read after being initialized in main.cc. so drop it.
Signed-off-by: Mykola Golub <mgolub@suse.com>
Signed-off-by: Kefu Chai <kchai@redhat.com>
We have ceph_fallocate, but we need the analogous call for the lowlevel
interface as well. Add ceph_ll_fallocate.
Tracker: http://tracker.ceph.com/issues/24893
Signed-off-by: Jeff Layton <jlayton@kernel.org>
handles the insights cases that had been using Ceph CLI
`erasure-code-profile ls/get` to access the same data.
Signed-off-by: Noah Watkins <nwatkins@redhat.com>
this addresses a warning introduced by f528475d:
src/tools/rbd/action/Create.cc:59:19: warning: the address of
'ConfigProxy& g_conf()' will never be NULL [-Waddress]
assert(g_conf != nullptr);
~~~~~~~^~~~
Signed-off-by: Kefu Chai <kchai@redhat.com>
The function is used for opening sessions for importing caps. It didn't
set client metadata for new sessions.
Signed-off-by: "Yan, Zheng" <zyan@redhat.com>
Info encapsulated in cd-info-card selector:
- Overall health status, MON/OSD/MDS/MGR status
Fixes: https://tracker.ceph.com/issues/24778
Signed-off-by: Alfonso Martínez <almartin@redhat.com>
unittest_seastar_echo cannot be run as unittest w/o starting two
instances.
unittest_seastar_messenger times out. need to RCA it.
Signed-off-by: Kefu Chai <kchai@redhat.com>
