For the MDS which needs to know about new blacklist entries
in order to evict corresponding MDS sessions.
Fixes: http://tracker.ceph.com/issues/17980
Signed-off-by: John Spray <john.spray@redhat.com>
we should set placement rule in RGWObjManifest::obj_iterator::update_location()
when ofs < manifest->get_head_size().
Signed-off-by: fang yuxiang fang.yuxiang@eisoo.com
as the latter is the only consumer of _gettermsize(). and a little bit
refactor to improve the readability and be more pep8 compliant.
Signed-off-by: Kefu Chai <kchai@redhat.com>
If the map has *no* choose_args it does not have incompat choose args.
This makes the OSDMap::get_features() check
if (crush->has_incompat_choose_args())
features |= CEPH_FEATURE_CRUSH_CHOOSE_ARGS;
happier.
Signed-off-by: Sage Weil <sage@redhat.com>
This could induce asserts in multi-nfs-gateway scenarios.
Fixes: http://tracker.ceph.com/issues/20047
Signed-off-by: Matt Benjamin <mbenjamin@redhat.com>
If a setattr (e.g., chown) has been performed on a bucket, then like
any directory it has a hidden object storing its attributes. This must
be deleted before attempting bucket delete, otherwise, actually empty
buckets will not be removable via NFS.
Fixes: http://tracker.ceph.com/issues/20045
Signed-off-by: Matt Benjamin <mbenjamin@redhat.com>
This will allow strict ADL to find the right overload in the case of a
container of containers and unbreak compilation with Clang.
Signed-off-by: Adam C. Emerson <aemerson@redhat.com>
We don't want to do the at-end.yaml scrubbing business with this test.
Move it into a separate collection until after luminous.
I have a todo item on the post-luminous cleanup list to avoid forgetting
to move this back.
Fixes: http://tracker.ceph.com/issues/19935
Signed-off-by: Sage Weil <sage@redhat.com>
- use 'noup' flag
- wait for healthy between each test
- check counts for each type
Fixes: http://tracker.ceph.com/issues/19990
Signed-off-by: Sage Weil <sage@redhat.com>
The old 'approaching max_size' code expects MDS set max_size to
'2 x reported_size'. This is no longer true. The new code reports
file size when half of previous max_size increment has been used.
Signed-off-by: "Yan, Zheng" <zyan@redhat.com>
For very large file, setting the writable range to '2 * file_size'
causes file recovery to run a long time. To recover a 1T file, Filer
needs to probe 2T~1T range.
Fixes: http://tracker.ceph.com/issues/19955
Signed-off-by: "Yan, Zheng" <zyan@redhat.com>
This allows for radosgw-admin to run even if there was a
misconfiguration in setting tier types which would otherwise make it
difficult to revert an invalid tier type setting. Since sync modules are
necessary only for rgw and not for the admin itself, this allows for
invalid tier configurations to be corrected as rgw-admin wouldn't fail
Fixes: http://tracker.ceph.com/issues/19830
Signed-off-by: Abhishek Lekshmanan <abhishek@suse.com>
rgw: fix lc list failure when shards not be all created
Reviewed-by: Daniel Gryniewicz <dang@redhat.com>
Reviewed-by: Casey Bodley <cbodley@redhat.com>
