# v12.2.6



* Backport #22637: luminous: rgw:lc: set LifecycleConfiguration without "Rule" tag return OK
* Backport #22696: luminous: client: dirty caps may never get the chance to flush
* Backport #22769: luminous: allow client requests to preempt scrub
* Backport #22934: luminous: filestore journal replay does not guard omap operations
* Backport #22937: luminous: beast: listen on multiple endpoints
* Backport #23151: luminous: doc: update ceph-fuse with FUSE options
* Backport #23157: luminous: mds: underwater dentry check in CDir::_omap_fetched is racy
* Backport #23227: luminous: clang compilation error in BoundedKeyCounter
* Backport #23231: luminous: rgw_statfs should report the correct stats
* Backport #23308: luminous: doc: Fix -d option in ceph-fuse doc
* Backport #23474: luminous: client: allow caller to request that setattr request be synchronous
* Backport #23475: luminous: ceph-fuse: trim ceph-fuse -V output
* Backport #23607: luminous: import-diff failed: (33) Numerical argument out of domain - if image size of the child is larger than the size of its parent
* Backport #23631: luminous: python bindings fixes and improvements
* Backport #23632: luminous: mds: handle client requests when mds is stopping
* Backport #23635: luminous: client: fix request send_to_auth was never really used
* Backport #23636: luminous: mds: kicked out by monitor during rejoin
* Backport #23637: luminous: mds: assertion in MDSRank::validate_sessions
* Backport #23638: luminous: ceph-fuse: getgroups failure causes exception
* Backport #23640: luminous: rbd: import with option --export-format fails to protect snapshot
* Backport #23666: luminous: SIGFPE, Arithmetic exception in AsyncConnection::_process_connection
* Backport #23668: luminous: There is no 'ceph osd pool get erasure allow_ec_overwrites' command
* Backport #23671: luminous: mds: MDBalancer using total (all time) request count in load statistics
* Backport #23672: luminous: bluestore: ENODATA on aio
* Backport #23675: luminous: qa/workunits/mon/test_mon_config_key.py fails on master
* Backport #23681: luminous: mg_read() call has wrong arguments
* Backport #23682: luminous: rgw:failed to pass test_bucket_list_maxkeys_unreadable in s3-test
* Backport #23683: luminous: radosgw-admin should not use metadata cache when not needed
* Backport #23685: luminous: rgw_file: post deadlock fix, direct deleted RGWFileHandle objects can remain in handle table
* Backport #23698: luminous: mds: load balancer fixes
* Backport #23700: luminous: osd: KernelDevice.cc: 539: FAILED assert(r == 0)
* Backport #23702: luminous: mds: sessions opened by journal replay do not get dirtied properly
* Backport #23703: luminous: MDSMonitor: mds health warnings printed in bad format
* Backport #23704: luminous: ceph-fuse: broken directory permission checking
* Backport #23750: luminous: mds: ceph.dir.rctime follows dir ctime not inode ctime
* Backport #23770: luminous: ceph-fuse: return proper exit code
* Backport #23771: luminous: client: fix gid_count check in UserPerm->deep_copy_from()
* Backport #23782: luminous: table of contents doesn't render for luminous/jewel docs
* Backport #23784: luminous: osd: Warn about objects with too many omap entries 
* Backport #23786: luminous: "utilities/env_librados.cc:175:33: error: unused parameter 'offset' [-Werror=unused-parameter]" in rados
* Backport #23791: luminous: MDSMonitor: new file systems are not initialized with the pending_fsmap epoch
* Backport #23792: luminous: MDSMonitor: MDSMonitor::encode_pending modifies fsmap not pending_fsmap
* Backport #23802: luminous: slow ceph_ll_sync_inode calls after setattr
* Backport #23808: luminous: upgrade: bad pg num and stale health status in mixed lumnious/mimic cluster
* Backport #23818: luminous: client: add option descriptions and review levels (e.g. LEVEL_DEV)
* Backport #23833: luminous: MDSMonitor: crash after assigning standby-replay daemon in multifs setup
* Backport #23835: luminous: mds: fix occasional dir rstat inconsistency between multi-MDSes
* Backport #23850: luminous: Read operations segfaulting multiple OSDs
* Backport #23852: luminous: OSD crashes on empty snapset
* Backport #23861: luminous: rgw: admin rest api shouldn't return error when getting user's stats if the user hasn't create any bucket.
* Backport #23862: luminous: aws4 auth not implemented for PutBucketRequestPayment
* Backport #23863: luminous: scrub interaction with HEAD boundaries and clones is broken
* Backport #23864: luminous: compression ratio depends on block size, which is much smaller (16K vs 4M) in multisite sync
* Backport #23865: luminous: [rgw] GET <object>?torrent returns object's body instead torrent-file
* Backport #23866: luminous: No meaningful error when RGW cannot create pools due to lack of available PGs
* Backport #23868: luminous: rgw: do not reflect period if not current in RGWPeriodPuller::pull
* Backport #23869: luminous: rgw sends garbage meta.compression to ElasticSearch
* Backport #23870: luminous: null map from OSDService::get_map in advance_pg
* Backport #23881: luminous: Bluestore OSD hit assert((log_reader->buf.pos & ~super.block_mask()) == 0)
* Backport #23886: luminous: Resharding hangs with versioning-enabled buckets
* Backport #23900: luminous: [rbd-mirror] asok hook for image replayer not re-registered after bootstrap
* Backport #23902: luminous: [rbd-mirror] local tag predecessor mirror uuid is incorrectly replaced with remote
* Backport #23904: luminous: Deleting a pool with active watch/notify linger ops can result in seg fault
* Backport #23906: luminous: libcurl ignores headers with empty value, leading to signature mismatches
* Backport #23912: luminous: mon: High MON cpu usage when cluster is changing
* Backport #23913: luminous: rbd-nbd can deadlock in logging thread
* Backport #23914: luminous: cache-try-flush hits wrlock, busy loops
* Backport #23915: luminous: monitors crashing ./include/interval_set.h: 355: FAILED assert(0) (jewel+kraken)
* Backport #23924: luminous: LibRadosAio.PoolQuotaPP failed
* Backport #23925: luminous: assert on pg upmap
* Backport #23930: luminous: mds: scrub code stuck at trimming log segments
* Backport #23931: luminous: qa: test_purge_queue_op_rate: self.assertTrue(phase2_ops < phase1_ops * 1.25)
* Backport #23933: luminous: client: avoid second lock on client_lock
* Backport #23934: luminous: client: "remove_session_caps still has dirty|flushing caps" when thrashing max_mds
* Backport #23935: luminous: mds: may send LOCK_SYNC_MIX message to starting MDS
* Backport #23936: luminous: cephfs-journal-tool: segfault during journal reset
* Backport #23945: luminous: potential race in rbd-mirror disconnect QA test
* Backport #23946: luminous: mds: crash when failover
* Backport #23950: luminous: mds: stopping rank 0 cannot shutdown until log is trimmed
* Backport #23951: luminous: mds: stuck during up:stopping
* Backport #23977: luminous: multisite: misleading error on normal shutdown
* Backport #23982: luminous: qa: TestVolumeClient.test_lifecycle needs updated for new eviction behavior
* Backport #23984: luminous: mds: scrub on fresh file system fails
* Backport #23985: luminous: librbd::Watcher's handle_rewatch_complete might fire after object destroyed
* Backport #23986: luminous: recursive lock of objecter session::lock on cancel
* Backport #23987: luminous: cephfs does not count st_nlink for directories correctly?
* Backport #23988: luminous: luminous->master: luminous crashes with AllReplicasRecovered in Started/Primary/Active/NotRecovering state
* Backport #23991: luminous: client: hangs on umount if it had an MDS session evicted
* Backport #24014: luminous: mgr/influx: Module fails to parse service names if RGW is present
* Backport #24015: luminous: UninitCondition in PG::RecoveryState::Incomplete::react(PG::AdvMap const&)
* Backport #24016: luminous: scrub interaction with HEAD boundaries and snapmapper repair is broken
* Backport #24042: luminous: ceph-disk log is written to /var/run/ceph
* Backport #24043: luminous: java compile error: Source/Target option 1.5 is not supported since jdk 9
* Backport #24048: luminous: pg-upmap cannot balance in some case
* Backport #24049: luminous: ceph-fuse: missing dentries in readdir result
* Backport #24050: luminous: mds: MClientCaps should carry inode's dirstat
* Backport #24055: luminous: VolumeClient: allow ceph_volume_client to create 'volumes' without namespace isolation
* Backport #24059: luminous: Deleting a pool with active notify linger ops can result in seg fault
* Backport #24060: luminous: RGW Multi-site radosgw-admin sync status and radosgw-admin bucket sync status commands addition of 'detail' flag
* Backport #24063: luminous: Misnamed S3 operation
* Backport #24070: luminous: build-integration-branch script can fail with UnicodeEncodeError
* Backport #24084: luminous: [rbd-mirror] bootstrap should not raise -EREMOTEIO if local image still attached
* Backport #24086: luminous: [rbd-mirror] potential races during PoolReplayer shut-down
* Backport #24107: luminous: PurgeQueue::_consume() could return true when there were no purge queue item actually executed.
* Backport #24108: luminous: MDCache.cc: 5317: FAILED assert(mds->is_rejoin())
* Backport #24120: luminous: rgw: 403 error when creating an object with metadata containing sequence of spaces
* Backport #24122: luminous: selinux denials with ceph-deploy/ceph-volume lvm device
* Backport #24130: luminous: mds: race with new session from connection and imported session
* Backport #24132: luminous: "122 - unittest_bluefs (OTHER_FAULT)" during ctest run
* Backport #24153: luminous: Eviction still raced with scrub due to preemption
* Backport #24156: luminous: [rbd-mirror] potential deadlock when running asok 'flush' command
* Backport #24185: luminous: client: segfault in trim_caps
* Backport #24188: luminous: kceph: umount on evicted client blocks forever
* Backport #24189: luminous: qa: kernel_mount.py umount must handle timeout arg
* Backport #24198: luminous: mon: slow op on log message
* Backport #24201: luminous: client: fails to respond cap revoke from non-auth mds
* Backport #24205: luminous: mds: broadcast quota to relevant clients when quota is explicitly set
* Backport #24207: luminous: client: deleted inode's Bufferhead which was in STATE::Tx would lead a assert fail
* Backport #24214: luminous: Module 'balancer' has failed: could not find bucket -14
* Backport #24216: luminous: "process (unknown)" in ceph logs
* Bug #24225: AArch64 CRC32 crash with SIGILL
* Backport #24245: luminous: Manager daemon y is unresponsive during teuthology cluster teardown
* Backport #24247: luminous: SharedBlob::put() racy
* Backport #24252: luminous: Admin OPS Api overwrites email when user is modified
* Backport #24258: luminous: crush device class: Monitor Crash when moving Bucket into Default root
* Backport #24279: luminous: RocksDB compression is not supported at least on Debian.
* Backport #24290: luminous: common: JSON output from rados bench write has typo in max_latency key
* Backport #24298: luminous: rgw: fix 'copy part' without 'x-amz-copy-source-range' when compression enabled
* Backport #24302: luminous: rgw: (jewel) can't delete swift acls with swift command.
* Backport #24314: luminous: multisite test failures in test_versioned_object_incremental_sync
* Backport #24328: luminous: assert manager.get_num_active_clean() == pg_num on rados/singleton/all/max-pg-per-osd.from-primary.yaml
* Backport #24331: luminous: mon: mds health metrics sent to cluster log indpeendently
* Backport #24334: luminous: ARMv8 feature detection broken, leading to illegal instruction crashes
* Backport #24341: luminous: mds memory leak
* Backport #24351: luminous: slow mon ops from osd_failure
* Backport #24353: luminous: rgw:  request with range defined as "bytes=0--1" returns 416 InvalidRange
* Backport #24356: luminous: osd: pg hard limit too easy to hit
* Bug #24369: luminous: checking quota while holding cap ref may deadlock
* Bug #24370: luminous: root dir's new snapshot lost when restart mds
* Backport #24374: luminous: mon: auto compaction on rocksdb should kick in more often
* Backport #24378: luminous: [rbd-mirror] entries_behind_master will not be zero after mirror over
* Backport #24393: luminous: rgw: making implicit_tenants backwards compatible
* Bug #24421: async messager thread cpu high, osd service not normal until restart 
* Backport #24477: luminous: Bucket lifecycles stick around after buckets are deleted
* Backport #24503: luminous: ObjectStore/StoreTestSpecificAUSize.SyntheticMatrixNoCsum/2 failed fsck with stray objects
* Bug #24511: osd crushed at thread_name:safe_timer
* Bug #24622: No module named rados
* Bug #24661: os/bluestore: don't store/use path_block.{db,wal} from meta 
* Documentation #24712: Memory recommendations for bluestore
* Backport #24740: luminous: CLI unit formatting tests are broken
* Backport #24750: luminous: os/bluestore: don't store/use path_block.{db,wal} from meta 
* Bug #24789: [rgw] ERROR: unable to remove bucket(2) No such file or directory (if --bypass-gc)
* Backport #24806: luminous: rgw workload makes osd memory explode
* Bug #24947: Ceph Luminous radosgw: Couldn't init storage provider (RADOS)
* Bug #25098: Bluestore OSD failed to start with `bluefs_types.h: 54: FAILED assert(pos <= end)`