Fixed:
** CID 1395347 (#1 of 1): Argument cannot be negative (NEGATIVE_RETURNS)
3. negative_returns: fd is passed to a parameter that cannot be negative
Signed-off-by: Amit Kumar amitkuma@redhat.com
The _process methods assumes the shard osdmap refs are
non-null, but that leaves a window between startup and
the first call to prune_pg_waiters (via consume_map()),
which can lead to a seg fault.
Calling prune during startup, before we have any work,
will initialize those references.
Fixes: http://tracker.ceph.com/issues/20748
Signed-off-by: Sage Weil <sage@redhat.com>
so we can avoid the warnings like
grep: Unmatched ( or \(
because we pass the whitelisted string to `egrep -v "$1"` directly.
Signed-off-by: Kefu Chai <kchai@redhat.com>
Well, data is certainly unavailable, and may also be
degraded in the sense that we can't peer. I think
unavailable is the more severe of the two, though, so
let's put it there!
Signed-off-by: Sage Weil <sage@redhat.com>
First paragraph: explain what the error means.
Second or later paragraph: describe steps to fix or mitigate.
Signed-off-by: Sage Weil <sage@redhat.com>
- s/cephfs_data/cephfs_data_a
- s/cephfs_metadata/cephfs_metadata_a
- s#./rados df#bin/rados df
- update the 'bin/rados df' output
- remove the rbd pool, it isn't created by default anymore.
Signed-off-by: Zhu Shangzhong <zhu.shangzhong@zte.com.cn>
msg/async: fix the bug of inaccurate calculation of l_msgr_send_bytes
Reviewed-by: Pan Liu <wanjun.lp@alibaba-inc.com>
Reviewed-by: Jos Collin <jcollin@redhat.com>
Reviewed-by: Haomai Wang <haomai@xsky.com>
- radosgw/s3/bucketops.rst: fix Malformed table.
- operations/health-checks.rst: Title underline too short
- rbd/rados-rbd-cmds.rst: Title underline too short
- rados/operations/index.rst: include health-checks in toc
Signed-off-by: Kefu Chai <kchai@redhat.com>
This does several little things that add up to big concurrency and safety
improvements:
* Switch to passing around PGRefs instead of raw pointers, which is
generally a good idea
* drop the pg_map_lock once we're done looking up the PGRefs, since
we don't need it and holding the PG pointer alive was the only previous
thing that might have made it necessary
* don't hold the recovery_lock since we don't need any OSD-level
synchronization
* make sure the PG is not being deleted before we do a force-change of its
state
Fixes: http://tracker.ceph.com/issues/20808
Signed-off-by: Greg Farnum <gfarnum@redhat.com>
This avoids crashing when older monitors do not support it.
Fixes: http://tracker.ceph.com/issues/20850
Signed-off-by: Josh Durgin <jdurgin@redhat.com>
The cluster is expected to become degraded during reboot.
Fixes: http://tracker.ceph.com/issues/20731
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
It seems many are initially unclear as to how the current
implementations of mclock op queues work, so we need to document it to
avoid confusion.
Signed-off-by: J. Eric Ivancich <ivancich@redhat.com>
Get rid of the undefined behavior of destroying condition variables
while they're being waited on.
Signed-off-by: Adam C. Emerson <aemerson@redhat.com>
The summary field can be extended with the fields
used to construct the message (e.g. including
the down osd count in the message about osds
being down).
The detail entries, similarly, can be extended
with machine-readable fields like the PG ID
for a damaged PG.
For the moment all the internal stuff is just
strings still, but we change the output format
so that we don't break it later when we
add things.
Signed-off-by: John Spray <john.spray@redhat.com>
This was working for setting values, but failing to call
the config observers, so some values didn't take effect.
Fixes: http://tracker.ceph.com/issues/20803
Signed-off-by: John Spray <john.spray@redhat.com>
