If sudo requires a password it is nice to be able to avoid
sudo sysctl by setting core_pattern ahead of time.
Signed-off-by: David Zafman <dzafman@redhat.com>
Up until version 1.1.3 of Sphinx, this call would show hidden toctree
entries. Most of Ceph's toctree's are hidden, so when the version of
Sphinx was bumped to >1.6, the sidebar stopped showing entries
Signed-off-by: Alfredo Deza <adeza@redhat.com>
Originally bluestore defines the two lat deferred_queued_lat and
deferred_aio_wait_lat, but doesn't use it.
This patch is to add the records that each TXC stays on deferred queue
deferred ios.
Signed-off-by: Xiaoyan Li xiaoyan.li@intel.com
This patch can solve the write issue when we use spdk nvme driver
to handle I/Os from db bdev in the future.
Signed-off-by: Ziye Yang <optimistyzy@gmail.com>
Signed-off-by: Pan Liu <wanjun.lp@alibaba-inc.com>
EC could set one or more members of acting set to CRUSH_ITEM_NONE,
which as a result can cause pgs_by_osd.resize() attempt to apply
for a large amount of memory which we can not afford.
Fix the above problem by always excluding a current DNE osd from
acting set.
Fixes: http://tracker.ceph.com/issues/20970
Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
os/bluestore: move aio_callback{,_priv} to base class BlockDevice
Reviewed-by: Pan Liu <liupan1111@gmail.com>
Reviewed-by: Kefu Chai <kchai@redhat.com>
Drop the unnecessary polling_stop() in the ~RDMAStack(). The RDMADispatcher object contained in
RDMAStack will get distroyed when RDMAStack gets distroyed. Thus ~RDMADispatcher() will be
called automatically and executes the polling_stop().
Fixes:
The below coverity scan reports.
1. CID 1416590: Memory - corruptions
2. CID 1416593: Memory - illegal accesses
3. CID 1416595: Memory - corruptions
4. CID 1416597: Memory - illegal accesses
Signed-off-by: Jos Collin <jcollin@redhat.com>
rbd-mirror: Set the data pool correctly when creating images
Reviewed-by: Mykola Golub <mgolub@mirantis.com>
Reviewed-by: Jason Dillaman <dillaman@redhat.com>
Return osd_min_pg_log_entries to its original values, which matches
osd_pg_log_dups_tracked, so the extended dup log is not used in the
general case.
This helps address: http://tracker.ceph.com/issues/21026
Signed-off-by: J. Eric Ivancich <ivancich@redhat.com>
Previously the data-pool option was ignored by rbd-mirror, so when using
erasure coding data for the image on the secondary site would end up in
the metadata pool.
Configure the data-pool using the data-pool's name from the primary site
when creating images on the secondary site.
Fixes: http://tracker.ceph.com/issues/20567
Signed-off-by: Adam Wolfe Gordon <awg@digitalocean.com>
(cherry picked from commit dffaac83d1)
rpm expands all macros in a .spec file, even those in comments. Drop the
percent signs so rpm will not expand these.
This change silences rpmlint's warning about macros in comments.
Signed-off-by: Ken Dreyer <kdreyer@redhat.com>
Fixes the coverity issue:
** 717332 Uninitialized pointer field
2. uninit_member: Non-static class member field iocb.data is not
initialized in this constructor nor in any functions that it calls.
4. uninit_member: Non-static class member field iocb.key is not
initialized in this constructor nor in any functions that it calls.
6. uninit_member: Non-static class member field iocb.__pad2 is not
initialized in this constructor nor in any functions that it calls.
8. uninit_member: Non-static class member field iocb.aio_lio_opcode
is not initialized in this constructor nor in any functions that it calls.
10. uninit_member: Non-static class member field iocb.aio_reqprio
is not initialized in this constructor nor in any functions that it calls.
CID 717332 (#1 of 1): Uninitialized pointer field (UNINIT_CTOR)
12. uninit_member: Non-static class member field iocb.aio_fildes is not
initialized in this constructor nor in any functions that it calls.
Signed-off-by: Amit Kumar amitkuma@redhat.com
if init fails to connect to a monitor, we just exit(1), there is no need
to set the `r` or lock the grand osd_lock for this. this change also
silences the warning of clang analyzer:
Value stored to 'r' is never read
Signed-off-by: Kefu Chai <kchai@redhat.com>
Support Filter tag in Lifecycle XML similar to AWS S3, while S3 docs
mention that this tag is mandatory, older clients still default to
Prefix, and S3 itself seems to relaxed in enforcing the rule, this
implementation also follows a similar pattern. Filter optionally
supports filtering via (multiple) Object Tags, this is still a TODO. The
current implementation of object tags is still as an object xattr, and
since the LC processing still iterates over the bucket index which
currently doesn't have any info. on tags, this requires some thought
into for implementing without a larger performance penalty
Fixes: http://tracker.ceph.com/issues/20872
Signed-off-by: Abhishek Lekshmanan <abhishek@suse.com>
