# v12.2.13



* Bug #23147: RGW: metrics 'qlen', 'qactive' are not work
* Backport #23223: luminous: rgw: garbage collector removes objects slowly
* Backport #23237: Corrupted downloads from civetweb when using multipart with slow connections
* Backport #24360: luminous: osd: leaked Session on osd.7
* Backport #36080: luminous: aarch64: Compiler-based detection of crc32 extended CPU type is broken
* Backport #37497: luminous: get or set rgw realm zonegroup zone should check user's caps for security
* Backport #37612: luminous: rpm: missing dependency on python34-ceph-argparse from python34-cephfs (and others?)
* Backport #37692: luminous: Image mirroring should be disabled when it is moved to trash
* Backport #37748: luminous: Add clear-data-digest command to objectstore tool
* Backport #37892: luminous: doc: wrong value of usage log default in logging section 
* Backport #38205: luminous: osds allows to partially start more than N+2
* Backport #38242: luminous: msg/async: connection race + winner fault can leave connection in standby
* Backport #38276: luminous: osd_map_message_max default is too high?
* Backport #38340: luminous: mds: may leak gather during cache drop
* Backport #38397: luminous: rgw: when exclusive lock fails due existing lock, log add'l info
* Backport #38436: luminous: crc cache should be invalidated when posting preallocated rx buffers
* Backport #38440: luminous: compare-and-write skips compare after copyup without object map
* Backport #38442: luminous: osd-markdown.sh can fail with CLI_DUP_COMMAND=1
* Backport #38445: luminous: mds: drop cache does not timeout as expected
* Backport #38508: luminous: [rbd-mirror] LeaderWatcher stuck in loop if pool deleted
* Backport #38551: luminous: core: lazy omap stat collection
* Backport #38564: luminous: [librbd] race condition possible when validating RBD pool
* Backport #38567: luminous: osd_recovery_priority is not documented (but osd_recovery_op_priority is)
* Backport #38674: luminous: Performance improvements for object-map
* Backport #38686: luminous: kcephfs TestClientLimits.test_client_pin fails with "client caps fell below min"
* Backport #38714: luminous: rgw: gc entries with zero-length chains are not cleaned up
* Backport #38719: luminous: crush: choose_args array size mis-sized when weight-sets are enabled
* Backport #38748: luminous: non existant mdlog failures logged at level 0
* Backport #38750: luminous: should report EINVAL in ErasureCode::parse() if m<=0
* Backport #38781: luminous: mgr/balancer: blame if upmap won't actually work
* Backport #38873: luminous: Rados.get_fsid() returning bytes in python3
* Backport #38877: luminous: mds: high debug logging with many subtrees is slow
* Backport #38880: luminous: ENOENT in collection_move_rename on EC backfill target
* Backport #38884: luminous: Lifecycle doesn't remove delete markers
* Backport #38887: luminous: GetBucketCORS API returns "Not Found" error code when CORS configuration does not exist
* Backport #38902: luminous: Minor rados related documentation fixes
* Backport #38905: luminous: osd/PGLog.h: print olog_can_rollback_to before deciding to rollback
* Backport #38908: luminous: rgw: read not exists null version success and return empty data
* Backport #38920: luminous: "Caught signal (Aborted) thread_name:radosgw" in ceph dashboard tests Jenkins job
* Backport #38925: luminous: beast frontend option to set the TCP_NODELAY socket option
* Backport #38954: luminous: backport krbd discard qa fixes to stable branches
* Backport #38958: luminous: multisite: sync status on master zone does not show "oldest incremental change not applied"
* Backport #38962: luminous: DaemonServer::handle_conf_change - broken locking
* Backport #38975: luminous: return ETIMEDOUT if we meet a timeout in poll
* Backport #39016: luminous: unable to cancel reshard operations for buckets with tenants
* Backport #39042: luminous: osd/PGLog: preserve original_crt to check rollbackability
* Backport #39177: luminous: rgw: remove_olh_pending_entries() does not limit the number of xattrs to remove
* Backport #39180: luminous: rgw: orphans find perf improvments
* Backport #39191: luminous: mds: crash during mds restart
* Backport #39198: luminous: mds: we encountered "No space left on device" when moving huge number of files into one directory
* Backport #39204: luminous: osd: leaked pg refs on shutdown
* Backport #39208: luminous: mds: mds_cap_revoke_eviction_timeout is not used to initialize Server::cap_revoke_eviction_timeout
* Backport #39213: luminous: mds: there is an assertion when calling Beacon::shutdown()
* Backport #39218: luminous: osd: FAILED ceph_assert(attrs || !pg_log.get_missing().is_missing(soid) || (it_objects != pg_log.get_log().objects.end() && it_objects->second->op == pg_log_entry_t::LOST_REVERT)) in PrimaryLogPG::get_object_context()
* Backport #39221: luminous: mds: behind on trimming and "[dentry] was purgeable but no longer is!"
* Backport #39227: luminous: rgw_file: can't retrieve etag of empty object written through NFS
* Backport #39231: luminous: kclient: nofail option not supported
* Backport #39239: luminous: "sudo yum -y install python34-cephfs" fails on mimic
* Backport #39243: luminous: msg/async: connection race + winner fault can leave connection stuck at replacing forever
* Backport #39247: luminous: os/bluestore: fix length overflow
* Backport #39254: luminous: occaionsal ObjectStore/StoreTestSpecificAUSize.Many4KWritesTest/2 failure
* Backport #39272: luminous: rgw: S3 policy evaluated incorrectly
* Backport #39277: luminous:  platform.linux_distribution() is deprecated; stop using it
* Backport #39314: luminous: krbd: fix rbd map hang due to udev return subsystem unordered
* Backport #39332: luminous: Build with lttng on openSUSE
* Backport #39343: luminous: ceph-objectstore-tool rename dump-import to dump-export
* Backport #39358: luminous: Compliance to aws s3's relaxed query handling behaviour
* Backport #39360: luminous: rgw:failed to pass test_bucket_create_naming_bad_punctuation in s3test
* Backport #39373: luminous: ceph tell osd.xx bench help : gives wrong help
* Bug #39395: ceph: ceph fs auth fails
* Backport #39409: luminous: inefficient unordered bucket listing
* Backport #39420: luminous: Don't mark removed osds in when running "ceph osd in any|all|*"
* Backport #39424: luminous: mgr: deadlock
* Backport #39427: luminous: 'rbd mirror status --verbose' will occasionally seg fault
* Backport #39431: luminous: Degraded PG does not discover remapped data on originating OSD
* Backport #39444: luminous: OSD crashed in BitmapAllocator::init_add_free()
* Backport #39457: luminous: mgr/prometheus: replace whitespaces in metric names
* Backport #39460: luminous: [rbd-mirror] "bad crc in data" error when listing large pools
* Backport #39463: luminous: print client IP in default debug_ms log level when "bad crc in {front|middle|data}" occurs
* Backport #39468: luminous:  There is no punctuation mark or blank between tid and client_id in the output of "ceph health detail" 
* Backport #39474: luminous: segv in fgets() in collect_sys_info reading /proc/cpuinfo
* Backport #39497: luminous: rgw admin: object stat command output's delete_at not readable
* Backport #39537: luminous: osd/ReplicatedBackend.cc: 1321: FAILED assert(get_parent()->get_log().get_log().objects.count(soid) && (get_parent()->get_log().get_log().objects.find(soid)->second->op == pg_log_entry_t::LOST_REVERT) && (get_parent()->get_log().get_log().object
* Backport #39563: luminous: Error message displayed when mon_osd_max_split_count would be exceeded is not as user-friendly as it could be
* Backport #39565: luminous: ceph-bluestore-tool: bluefs-bdev-expand silently bypasses main device (slot 2)
* Backport #39572: luminous: send x-amz-version-id header in PUT response
* Backport #39589: luminous: qa/tasks/rbd_fio: fixed missing delimiter between 'cd' and 'configure' 
* Backport #39603: luminous: document CreateBucketConfiguration for s3 PUT Bucket request
* Backport #39615: luminous: civetweb frontend: response is buffered in memory if content length is not explicitly specified
* Backport #39638: luminous: fsck on mkfs breaks ObjectStore/StoreTestSpecificAUSize.BlobReuseOnOverwrite
* Backport #39673: luminous: [test] possible race condition in rbd-nbd disconnect
* Backport #39691: luminous: mds:  error "No space left on device" when create a large number of dirs
* Backport #39696: luminous: rgw: success returned for put bucket versioning on a non existant bucket
* Backport #39719: luminous: short pg log+nautilus-p2p-stress-split: "Error: finished tid 3 when last_acked_tid was 5" in upgrade:nautilus-p2p
* Backport #39727: luminous: [test] devstack is broken (again)
* Backport #39732: luminous: rgw: allow radosgw-admin bucket list to use the --allow-unordered flag
* Backport #39733: luminous: multisite: mismatch of bucket creation times from List Buckets
* Backport #39747: luminous: Add support for --bypass-gc flag of radosgw-admin bucket rm command in RGW Multi-site
* Backport #40004: luminous: do_cmake.sh: "source" not found
* Backport #40032: luminous: rgw metadata search (elastic search): meta sync: ERROR: failed to read mdlog info with (2) No such file or directory
* Backport #40041: luminous: avoid trimming too many log segments after mds failover
* Backport #40082: luminous: osd: Better error message when OSD count is less than osd_pool_default_size
* Backport #40092: luminous: Missing Documentation for radosgw-admin reshard commands (man pages)
* Backport #40127: luminous: rgw: Swift interface: server side copy fails if object name contains `?`
* Backport #40132: luminous: rgw: putting X-Object-Manifest via TempURL should be prohibited
* Backport #40135: luminous: rgw: the Multi-Object Delete operation of S3 API wrongly handles the "Code" response element
* Backport #40138: luminous: document steps to disable metadata_heap on existing zones
* Backport #40143: luminous: multisite: 'radosgw-admin bucket sync status' should call syncs_from(source.name) instead of id
* Backport #40149: luminous: rgw: bucket may redundantly list keys after BI_PREFIX_CHAR
* Backport #40160: luminous: FSAL_CEPH assertion failed in Client::_lookup_name: "parent->is_dir()"
* Backport #40163: luminous: mount: key parsing fail when doing a remount
* Backport #40166: luminous: client: ceph.dir.rctime xattr value incorrectly prefixes "09" to the nanoseconds component 
* Bug #40182: luminous: pybind: luminous volume client breaks against nautilus cluster
* Bug #40200: luminous: mds: does fails assert(session->get_nref() == 1) when balancing
* Backport #40218: luminous: TestMisc.test_evict_client fails
* Backport #40221: luminous: mds: reset heartbeat during long-running loops in recovery
* Backport #40229: luminous: maybe_remove_pg_upmap can be super inefficient for large clusters
* Backport #40233: luminous: [CLI]rbd: get positional argument error when using --image
* Backport #40266: luminous: data race in OutputDataSocket
* Bug #40286: luminous: qa: remove ubuntu 14.04 testing
* Backport #40318: luminous: "make: *** [hello_world_cpp] Error 127" in rados
* Backport #40343: luminous: mds: fix corner case of replaying open sessions
* Backport #40347: luminous: ssl tests failing with SSLError("bad handshake: Error([('SSL routines', 'tls_process_server_certificate', 'certificate verify failed'),)",)
* Backport #40350: luminous: rgw/OutputDataSocket: append_output(buffer::list&) says it will (but does not) discard output at data_max_backlog 
* Backport #40359: luminous: rgw: set null version object issues
* Backport #40422: luminous: Bitmap allocator return duplicate entries which cause interval_set assert
* Backport #40463: luminous: possible crash when replaying journal with invalid/corrupted ranges
* Backport #40496: luminous: Object Gateway multisite document read-only argument error
* Backport #40499: luminous: [cli] 'export' should handle concurrent IO completions
* Backport #40502: luminous: osd: rollforward may need to mark pglog dirty
* Backport #40506: luminous: rgw: conditionally allow builtin users with non-unique email addresses
* Backport #40534: luminous: pool compression options not consistently applied
* Backport #40548: luminous: Keyrings created by ceph auth get are not suitable for ceph auth import
* Backport #40551: luminous: [test] qemu-iotests tests fails under latest Ubuntu kernel
* Backport #40559: luminous: rgw: the log output gets very spammy in multisite clusters
* Backport #40574: luminous: Disabling journal might result in assertion failure
* Bug #40584: kernel build failure in kernel_untar_build.sh
* Backport #40592: luminous: rbd_mirror/ImageSyncThrottler.cc: 61: FAILED ceph_assert(m_queue.empty())
* Backport #40638: luminous: osd: report omap/data/metadata usage
* Backport #40650: luminous: os/bluestore: fix >2GB writes
* Backport #40653: luminous: Lower the default value of osd_deep_scrub_large_omap_object_key_threshold
* Backport #40697: luminous: test_envlibrados_for_rocksdb.yaml fails installing g++-4.7 on 18.04
* Backport #40735: luminous: multisite: failover docs should use 'realm pull' instead of 'period pull'
* Backport #40756: luminous: stupid allocator might return extents with length = 0
* Backport #40807: luminous: mds: msg weren't destroyed before handle_client_reconnect returned, if the reconnect msg was from non-existent session
* Backport #40852: luminous: multisite: radosgw-admin commands should not modify metadata on a non-master zone
* Backport #40880: luminous: Reduce log level for cls/journal and cls/rbd expected errors
* Backport #40892: luminous: mds: cleanup truncating inodes when standby replay mds trim log segments
* Backport #40947: luminous: Better default value for osd_snap_trim_sleep
* Backport #40978: luminous: missing string substitution when reporting mounts
* Backport #41000: luminous: client: failed to drop dn and release caps causing mds stary stacking.
* Backport #41020: luminous: simple: when 'type' file is not present activate fails
* Backport #41057: luminous: ceph-volume does not recognize wal/db partitions created by ceph-disk
* Backport #41104: luminous: rgw: when usring radosgw-admin to list bucket, can set --max-entries excessively high
* Backport #41111: luminous: rgw: fix drain handles error when deleting bucket with bypass-gc option
* Backport #41139: luminous: ceph-volume prints errors to stdout with --format json
* Backport #41202: luminous: ceph-volume prints log messages to stdout
* Backport #41266: luminous: beast frontend throws an exception when running out of FDs
* Backport #41278: luminous: mgr/prometheus:  Setting scrape_interval breaks cache timeout comparison
* Backport #41281: luminous: BlueStore tool to check fragmentation
* Backport #41285: luminous: error from replay does not stored in rbd-mirror status
* Backport #41289: luminous: fix and improve doc regarding manual bluestore cache settings.
* Backport #41322: luminous: multisite: datalog/mdlog trim don't loop until done
* Backport #41334: luminous: ceph-test RPM not built for SUSE
* Backport #41338: luminous: os/bluestore/BlueFS: use 64K alloc_size on the shared device
* Bug #41367: rocksdb: submit_transaction error: Corruption: block checksum mismatch code = 2
* Bug #41370: [RGW] RGW in website mode: rgw_rados.h: 2150: FAILED assert(!obj.empty()
* Backport #41373: luminous: batch functional idempotency test fails since message is now on stderr
* Backport #41382: luminous: rgw: housekeeping of reset stats operation in radosgw-admin and cls back-end
* Bug #41401: rgw: api_name fixes from Nautilus (e.g., allows CreateBucket w/alternate placement)
* Backport #41421: luminous: `rbd mirror pool status --verbose` test is missing
* Backport #41439: luminous: [rbd-mirror] cannot connect to remote cluster when running as 'ceph' user
* Backport #41458: luminous: proc_replica_log need preserve replica log's crt
* Backport #41480: luminous: rgw dns name is not case sensitive
* Backport #41489: luminous: client: client should return EIO when it's unsafe reqs have been dropped when the session is close.
* Backport #41510: luminous: 50-100% iops lost due to bluefs_preextend_wal_files = false
* Backport #41532: luminous: Move bluefs alloc size initialization log message to log level 1
* Backport #41544: luminous: [test] rbd-nbd FSX test runs are failing
* Backport #41579: luminous: rgw: api_name fixes from Nautilus (e.g., allows CreateBucket w/alternate placement)
* Backport #41613: luminous: ceph-volume lvm list is O(n^2)
* Backport #41621: luminous: in rbd-ggate the assert in Log:open() will trigger
* Backport #41626: luminous: multisite: ENOENT errors from FetchRemoteObj causing bucket sync to stall without retry
* Backport #41644: luminous: QA run failures "Command failed on smithi with status 1: '\n sudo yum -y install ceph-radosgw\n ' "
* Backport #41697: luminous: Network ping monitoring
* Backport #41706: luminous: in cls_bucket_list_unordered() listing of entries following an entry for which check_disk_state() returns -ENOENT may not get listed
* Backport #41709: luminous: Set concurrent max_background_compactions in rocksdb to 2
* Backport #41713: luminous: can't remove rados objects after copy rgw-object fail
* Backport #41730: luminous: osd/ReplicatedBackend.cc: 1349: FAILED ceph_assert(peer_missing.count(fromshard))
* Backport #41733: luminous: osd: need clear PG_STATE_CLEAN when repair object
* Backport #41772: luminous: RBD image manipulation using python API crashing since Nautilus
* Backport #41808: luminous: rgw: fix minimum of unordered bucket listing
* Backport #41845: luminous: tools/rados: allow list objects in a specific pg in a pool
* Backport #41864: luminous: Mimic MONs have slow/long running ops
* Backport #41914: luminous: mgr/test_localpool.sh fails after multiple tries on luminous
* Backport #41919: luminous: osd: scrub error on big objects; make bluestore refuse to start on big objects
* Backport #41959: luminous: tools/rados: add --pgid in help
* Backport #41962: luminous: Segmentation fault in rados ls when using --pgid and --pool/-p together as options
* Backport #42037: luminous: Enable auto-scaler and get src/osd/PeeringState.cc:3671: failed assert info.last_complete == info.last_update
* Backport #42039: luminous: client: _readdir_cache_cb() may use the readdir_cache already clear
* Backport #42049: luminous: fix pytest warnings
* Bug #42056: rgw: librgw write wrongly closed in NFS3
* Bug #42058: OSD reconnected across map epochs, inconsistent pg logs created
* Backport #42127: luminous: mgr/balancer FAILED ceph_assert(osd_weight.count(i.first))
* Backport #42138: luminous: Remove unused full and nearful output from OSDMap summary
* Backport #42153: luminous: Removed OSDs with outstanding peer failure reports crash the monitor
* Bug #42175:  _txc_add_transaction error (2) No such file or directory not handled on operation 15
* Bug #42193: luminous: MDS crash running upgrade test
* Backport #42199: luminous: osd/PrimaryLogPG.cc: 13068: FAILED ceph_assert(obc)
* Backport #42241: luminous: Adding Placement Group id in Large omap log message
* Backport #42264: luminous: mimic and luminous still need to read ceph.conf.template from teuthology
* Bug #42316: msg/async: do not bump connect_seq for fault during ACCEPTING_SESSION
* Backport #42361: luminous: python3-cephfs should provide python36-cephfs
* Backport #42390: luminous: mgr/balancer: 'dict_keys' object does not support indexing
* Backport #42393: luminous: CephContext::CephContextServiceThread might pause for 5 seconds at shutdown
* Backport #42415: luminous: sphinx spits warning when rendering doc/rbd/qemu-rbd.rst 
* Backport #42425: luminous: [rbd] rbd map hangs up infinitely after osd down
* Backport #42527: luminous: concurrent "rbd unmap" failures due to udev
* Backport #42548: luminous: verify_upmaps can not cancel invalid upmap_items in some cases
* Backport #42573: luminous: restful: Query nodes_by_id for items
* Backport #42580: luminous: p2p tests fail due to missing python3-cephfs package
* Backport #42586: luminous: out of order caused by letting old msg from down peer be processed to RESETSESSION
* Backport #42663: luminous: RBD mirroring test cases broken in mimic due to bad backport
* Backport #42672: luminous: qa: cfuse_workunit_kernel_untar_build fails on Ubuntu 18.04
* Backport #42678: luminous: qa: malformed job
* Backport #42698: luminous: Larger cluster using upmap mode balancer can block other balancer commands
* Backport #42774: luminous: mds: add command that modify session metadata
* Backport #42784: luminous: mgr/prometheus: UnboundLocalError occurs when obj_store is neither filestore nor bluestore
* Backport #42796: luminous: unnecessary error message "calc_pg_upmaps failed to build overfull/underfull"
* Bug #42828: rbd journal err assert(ictx->journal != __null)  when  release exclusive_lock
* Backport #42834: luminous: STATE_KV_SUBMITTED is set too early.
* Backport #42849: luminous: ceph osd status - units invisible using black background
* Backport #42895: luminous: rgw: add list user admin OP API
* Backport #42988: luminous: update kernel.sh for read-only changes
* Backport #43013: luminous: rgw: crypt: permit RGW-AUTO/default with SSE-S3 headers 
* Backport #43093: luminous: Improve OSDMap::calc_pg_upmaps() efficiency
* Bug #43175: pgs inconsistent, union_shard_errors=missing
* Backport #43234: luminous: rgw: radosgw_admin teuthology task: No module named bunch
* Bug #43269: rgw: lc: continue past get_obj_state() failure
* Backport #43278: luminous: "cd /home/ubuntu/cephtest/s3-tests && ./bootstrap" fails on ubuntu
* Backport #43325: luminous: wrong datatype describing crush_rule
* Bug #43421: mon spends too much time to build incremental osdmap
* Backport #43499: luminous: rbd-mirror daemons don't logrotate correctly
* Backport #43532: luminous: Change default upmap_max_deviation to 5
* Bug #43562: Error in tcmalloc
* Backport #43577: luminous: StupidAllocator.cc: 265: FAILED assert(intervals <= max_intervals)
* Backport #43651: luminous: Improve upmap change reporting in logs
* Backport #43759: luminous: functional tests only assume correct number is osds if branch tests is mimic or luminous
* Backport #43926: luminous: kernel_untar_build.sh: bison: command not found
* Bug #44008: multi-part upload will lost part data when you abort and resume a multipart upload request by using aws java Signature Version 4 api
* Bug #44967: rgw:rgw crash when putting object tagging and post object with malformedXML