Executed ceph-conf --dump-all on a freshly installed v14.2.2 (nautilus)
cluster. Compared the global defaults to the keys/values specified in
mon-config-ref.rst. Checked options.cc to make sure the obsolete keys
are no longer used.
Fixes: https://tracker.ceph.com/issues/41516
Signed-off-by: James McClune <jmcclune@mcclunetechnologies.net>
If we have to fabricate a merge target, we need to prime any future splits
it might have. Otherwise a sequence like
- e100 1.f merge to 1.7
- e110 1.7 split to 1.f, 1.17, 1.1f
where we process all of the above in one go at, say, e120, will lead to
a crash in register_and_wake_split_child because 1.17 and/or 1.1f aren't
primed.
Fix by making identify_splits_and_merges do a recursive scan on any
merge/split participants detected too.
Fixes: http://tracker.ceph.com/issues/38483
Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
When the refcount does a "get", the tag ends with a null character,
but during a rollback, when the tag is "put" it does not end with a
null character, so the same tag is not being manipulated and the
refcounts are therefore off, preventing future deletion of an rgw
object. This adds the null character to the "put" operation.
This change was originally submitted by lltlo <120680451@qq.com> but
wasn't signed-off.
Signed-off-by: J. Eric Ivancich <ivancich@redhat.com>
This change was originally submitted by lltlo <120680451@qq.com> but
had some formatting issues and wasn't signed-off.
Signed-off-by: J. Eric Ivancich <ivancich@redhat.com>
Even in the case the full path was not created, make sure mkdir
does create the full path.
And while there replaced some `test ....` with `[ ... ]`
like in the remainder of the file.
Signed-off-by: Willem Jan Withagen <wjw@digiware.nl>
This commit refactors the script to make it more user-friendly and
maintainable. Added features:
* script now determine release/milestone from backport issue instead of
requiring user to provide it
* script now generates a more verbose PR description
* backport PR title is based on original PR instead of the tracker issue
* improved error handling
* new --debug option that triggers "set -x"
* new --verbose option
* errors and diagnostic messages are now printed to stderr
Signed-off-by: Nathan Cutler <ncutler@suse.com>
backfill_toofull seen on cluster where the most full OSD is at 1%
Reviewed-by: xie xingguo <xie.xingguo@zte.com.cn>
Reviewed-by: Neha Ojha <nojha@redhat.com>
mgr/dashboard: Extract E2E wait calls into PageHelper
Reviewed-by: Patrick Seidensal <pseidensal@suse.com>
Reviewed-by: Stephan Müller <smueller@suse.com>
Reviewed-by: Laura Paduano <lpaduano@suse.com>
It turns out this test has not been triggering the right code path for
a while now (krbd_parent_overlap.t was added just over a year ago, but
the original test goes back a few years).
To do its job it needs an empty clone with the overlap set to somewhere
in the middle of an object. The test relied on shrinking for this, but
shrinking no longer fits the bill because a copyup is now performed for
the object with the overlap -- we never go to the parent after that.
Let's size the parent so that it ends in the middle of an object and
then grow an empty clone instead. Also, while reading from page cache
does seem to trigger a read that straddles the overlap, it's not very
reliable, so supplement it with a direct read.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
`socket.remote_endpoint()` can throw exceptions corresponding to errors in the
`getpeername` syscall, make sure these are handled.
Fixes: CVE-2019-10222, https://tracker.ceph.com/issues/40018
Signed-off-by: Abhishek Lekshmanan <abhishek@suse.com>
rgw: fix missing tenant prefix in bucket name during bucket link
Reviewed-by: Casey Bodley <cbodley@redhat.com>
Reviewed-by: J. Eric Ivancich <ivancich@redhat.com>
Before this commit, backport-resolve-issue would blindly grab the first tracker
URL it saw in the GitHub PR description. With this commit, it examines all the
tracker URLs it sees in the PR description and uses the first one that
corresponds to an issue in the Redmine Backport tracker.
Signed-off-by: Nathan Cutler <ncutler@suse.com>
otherwise it is likely to race with any follow-up scrubbing activities.
Note that it's ok for us to requeue_scrub before re-executing the op
since scrubbing will be blocking for the (same) global pg lock we
are currently holding.
Fixes: https://tracker.ceph.com/issues/41427
Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
We don't want to cache any object_contexts through the interval change.
We actually assert that all currently live references are dead by the
time the flush for the next interval completes.
On detecting a interval changing, let's cancel any in-flight manifest
ops, just as we do for any other cache-tier ops.
Fixes: https://tracker.ceph.com/issues/41514
Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
