# v19.0.0

03/01/2024

Squid dev

* Bug #24403: mon failed to return metadata for mds
* Bug #43221: rgw: GET Bucket fails on renamed bucket on archive zone
* Bug #43393: qa: add support/qa for cephfs-shell on CentOS 9 / RHEL9
* Bug #44565: src/mds/SimpleLock.h: 528: FAILED ceph_assert(state == LOCK_XLOCK || state == LOCK_XLOCKDONE || state == LOCK_XLOCKSNAP || state == LOCK_LOCK_XLOCK || state == LOCK_LOCK || is_locallock())
* Bug #44660: Multipart re-uploads cause orphan data
* Bug #44916: client: syncfs flush is only fast with a single MDS
* Bug #45736: rgw: lack of headers in 304 response
* Bug #48673: High memory usage on standby replay MDS
* Bug #48678: client: spins on tick interval
* Bug #48750: ceph config set using osd/host mask not working
* Fix #51177: pybind/mgr/volumes: investigate moving calls which may block on libcephfs into another thread
* Bug #51824: pacific scrub ~mds_dir causes stray related ceph_assert, abort and OOM
* Bug #52280: Mds crash and fails with assert on prepare_new_inode
* Documentation #52656: mgr/prometheus: wron unit in RBD latency metric description
* Bug #53192: High cephfs MDS latency and CPU load with snapshots and unlink operations
* Bug #53240: full-object read crc is mismatch, because truncate modify oi.size and forget to clear data_digest
* Bug #53724: mds: stray directories are not purged when all past parents are clear
* Bug #54182: OSD_TOO_MANY_REPAIRS cannot be cleared in >=Octopus
* Bug #54557: scrub repair does not clear earlier damage health status
* Bug #54833: crash: void Locker::handle_file_lock(ScatterLock*, ceph::cref_t<MLock>&): assert(lock->get_state() == LOCK_LOCK || lock->get_state() == LOCK_MIX || lock->get_state() == LOCK_MIX_SYNC2)
* Bug #55165: client: validate pool against pool ids as well as pool names
* Bug #55606: [ERR]   Unhandled exception from module ''devicehealth'' while running on mgr.y: unknown
* Feature #55940: quota: accept values in human readable format as well
* Bug #56011: fs/thrash: snaptest-snap-rm-cmp.sh fails in mds5sum comparison
* Bug #56034: qa/standalone/osd/divergent-priors.sh fails in test TEST_divergent_3()
* Bug #56192: crash: virtual Monitor::~Monitor(): assert(session_map.sessions.empty())
* Bug #56239: crash: File "mgr/devicehealth/module.py", in get_recent_device_metrics: return self._get_device_metrics(devid, min_sample=min_sample)
* Bug #56397: client: `df` will show incorrect disk size if the quota size is not aligned to 4MB
* Bug #56577: mds: client request may complete without queueing next replay request
* Bug #56695: [RHEL stock] pjd test failures(a bug that need to wait the unlink to finish)
* Bug #57048: osdc/Journaler: better handle ENOENT during replay as up:standby-replay
* Bug #57071: mds: consider mds_cap_revoke_eviction_timeout for get_late_revoking_clients()
* Bug #57087: qa: test_fragmented_injection (tasks.cephfs.test_data_scan.TestDataScan) failure
* Bug #57154: kernel/fuse client using ceph ID with uid restricted MDS caps cannot update caps
* Bug #57206: ceph_test_libcephfs_reclaim crashes during test
* Bug #57244: [WRN] : client.408214273 isn't responding to mclientcaps(revoke), ino 0x10000000003 pending pAsLsXsFs issued pAsLsXsFs, sent 62.303702 seconds ago
* Bug #57641: Ceph FS fscrypt clones missing fscrypt metadata
* Bug #57655: qa: fs:mixed-clients kernel_untar_build failure
* Bug #57985: mds: warning `clients failing to advance oldest client/flush tid` seen with some workloads
* Fix #58023: mds: do not evict clients if OSDs are laggy
* Feature #58057: cephfs-top: enhance fstop tests to cover testing displayed data
* Feature #58129: mon/FSCommands: support swapping file systems by name
* Feature #58154: mds: add minor segment boundaries
* Bug #58195: mgr/snap_schedule: catch all exceptions to avoid crashing module
* Bug #58228: mgr/nfs: disallow non-existent paths when creating export
* Bug #58303: active mgr crashes with segfault when running 'ceph osd purge'
* Bug #58340: mds: fsstress.sh hangs with multimds (deadlock between unlink and reintegrate straydn(rename))
* Bug #58394: nofail option in fstab not supported
* Bug #58411: mds: a few simple operations crash mds
* Bug #58482: mds: catch damage to CDentry's first member before persisting
* Feature #58550: mds: add perf counter to track (relatively) larger log events
* Bug #58569: Add the ability to configure options for ceph-volume to pass to cryptsetup
* Bug #58591: report "Insufficient space (<5GB)" even when disk size is sufficient
* Bug #58619: mds: client evict [-h|--help] evicts ALL clients
* Bug #58645: Unclear error when creating new subvolume when subvolumegroup has ceph.dir.subvolume attribute set to 1
* Bug #58677: cephfs-top: test the current python version is supported
* Feature #58680: libcephfs: clear the suid/sgid for fallocate
* Bug #58812: ceph-volume prepare doesn't use partitions as-is anymore
* Bug #58832: ceph-mgr package installation fails on centos 9
* Feature #58877: mgr/volumes: regenerate subvolume metadata for possibly broken subvolumes
* Bug #58924: mgr: block register_client on new MgrMap
* Bug #58943: ceph-volume's deactivate doesn't close encrypted volumes
* Cleanup #58961: mgr/dashboard: remove old dashboard (dashboard v3)
* Bug #58971: mon/MDSMonitor: do not trigger propose on error from prepare_update
* Bug #58972: mon/OSDMonitor: do not propose on error in prepare_update
* Cleanup #58973: mgr/dashboard: RGW 404 shouldn't trigger log exceptions
* Bug #58974: mon/MonmapMonitor: do not propose on error in prepare_update
* Bug #58975: mon: do not erroneously propose on error in ::prepare_update
* Bug #59042: mon/AuthMonitor: do not erroneously propose on error in ::prepare_update
* Bug #59043: mon/ConfigMonitor: do not erroneously propose on error in ::prepare_update
* Bug #59044: mon/HealthMonitor: do not erroneously propose on error in ::prepare_update
* Bug #59045: mon/KVMonitor: do not erroneously propose on error in ::prepare_update
* Bug #59046: mon/LogMonitor: do not erroneously propose on error in ::prepare_update
* Bug #59047: mon/MgrStatMonitor: do not erroneously propose on error in ::prepare_update
* Bug #59067: mds: add cap acquisition throttled event to MDR
* Bug #59107: MDS imported_inodes metric is not updated.
* Bug #59134: mds: deadlock during unlink with multimds (postgres)
* Bug #59183: cephfs-data-scan: does not scan_links for lost+found
* Bug #59185: MDSMonitor: should batch propose osdmap/mdsmap changes via some fs commands
* Bug #59188: cephfs-top: cephfs-top -d <seconds> not working as expected
* Bug #59192: cls/test_cls_sdk.sh: Health check failed: 1 pool(s) do not have an application enabled (POOL_APP_NOT_ENABLED)
* Bug #59230: Test failure: test_object_deletion (tasks.cephfs.test_damage.TestDamage)
* Bug #59297: qa: test_join_fs_unset failure
* Bug #59314: mon/MDSMonitor: plug PAXOS when evicting an MDS
* Bug #59318: mon/MDSMonitor: daemon booting may get failed if mon handles up:boot beacon twice
* Feature #59328:  mgr/dashboard: add support for editing RGW zone
* Bug #59332: qa: test_rebuild_simple checks status on wrong file system
* Bug #59343: qa: fs/snaps/snaptest-multiple-capsnaps.sh failed
* Bug #59344: qa: workunit test fs/quota/quota.sh failed with "setfattr: .: Invalid argument"
* Bug #59350: qa: test_flag_scrub_mdsdir (tasks.cephfs.test_scrub_checks.TestScrubChecks) ... ERROR
* Feature #59388: mds/MDSAuthCaps: "fsname", path, root_squash can't be in same cap with uid and/or gids
* Bug #59413: cephfs: qa snaptest-git-ceph.sh failed with "got remote process result: 128"
* Bug #59425: qa: RuntimeError: more than one file system available
* Bug #59463: mgr/nfs: Setting NFS export config using -i option is not working
* Bug #59514: client: read wild pointer when reconnect to mds
* Bug #59527: qa: run scrub post disaster recovery procedure
* Bug #59529: cluster upgrade stuck with OSDs and MDSs not upgraded.
* Bug #59551: mgr/stats: exception ValueError ：invalid literal for int() with base 16: '0x'
* Bug #59552: mon: block osd pool mksnap for fs pools
* Bug #59553: cephfs-top: fix help text for delay
* Bug #59569: mds: allow entries to be removed from lost+found directory
* Bug #59580: memory leak (RESTful module, maybe others?)
* Bug #59582: snap-schedule: allow retention spec to specify max number of snaps to retain
* Bug #59624: pybind/ceph_argparse: Error message is not descriptive for ceph tell command 
* Bug #59657: qa: test with postgres failed (deadlock between link and migrate straydn(rename))
* Fix #59667: qa: ignore cluster warning encountered in test_refuse_client_session_on_reconnect
* Bug #59682: CephFS: Debian cephfs-mirror package in the Ceph repo doesn't install the unit file or man page
* Bug #59683: Error: Unable to find a match: userspace-rcu-devel libedit-devel device-mapper-devel with fscrypt tests
* Bug #59684: Test failure: test_fscrypt_dummy_encryption_with_quick_group (tasks.cephfs.test_fscrypt.TestFscrypt)
* Bug #59691: mon/MDSMonitor: may lookup non-existent fs in current MDSMap
* Bug #59705: client: only wait for write MDS OPs when unmounting
* Feature #59714: mgr/volumes: Support to reject CephFS clones if cloner threads are not available
* Bug #59716: tools/cephfs/first-damage: unicode decode errors break iteration
* Feature #59727: The libradosstriper interface provides an optional parameter to avoid shared lock when reading data
* Bug #59735: fs/ceph: cross check passed in fsid during mount with cluster fsid
* Bug #59813: crash: void PaxosService::propose_pending(): assert(have_pending)
* Bug #61009: crash: void interval_set<T, C>::erase(T, T, std::function<bool(T, T)>) [with T = inodeno_t; C = std::map]: assert(p->first <= start)
* Bug #61148: dbench test results in call trace in dmesg
* Bug #61182: qa: workloads/cephfs-mirror-ha-workunit - stopping mirror daemon after the test finishes timesout.
* Bug #61186: mgr/nfs: hitting incomplete command returns same suggestion twice
* Bug #61201: qa: test_rebuild_moved_file (tasks/data-scan) fails because mds crashes in pacific
* Bug #61332: dbench test results in call trace in dmesg
* Bug #61369: [reef] RGW crashes when replication rules are set using PutBucketReplication S3 API
* Fix #61378: mds: turn off MDS balancer by default
* Bug #61399: qa: build failure for ior
* Bug #61400: valgrind+ceph-mon: segmentation fault in rocksdb+tcmalloc
* Bug #61409: qa: _test_stale_caps does not wait for file flush before stat
* Feature #61417: Zoned Block Devices (ZNS) support
* Bug #61444: mds: session ls command appears twice in command listing
* Bug #61459: mds: session in the importing state cannot be cleared if an export subtree task is interrupted while the state of importer is acking
* Bug #61466: Add bluefs write op count metrics
* Bug #61523: client: do not send metrics until the MDS rank is ready
* Bug #61572: mgr: remove invalid zero performance counter
* Bug #61574: qa: build failure for mdtest project
* Feature #61595: Consider setting "bulk" autoscale pool flag when automatically creating a data pool for CephFS
* Feature #61599: mon/MDSMonitor: optionally forbid to use standby for another fs as last resort
* Bug #61627: Mds crash and fails with assert on prepare_new_inode
* Bug #61629: rgw: add support http_date if http_x_amz_date is missing for sigv4
* Bug #61666: cephfs: print better error message when MDS caps perms are not right
* Bug #61690: mgr/dashboard: install_deps.sh fails on vanilla CentOS 8 Stream
* Bug #61749: mds/MDSRank: op_tracker of mds have slow op alway.
* Bug #61753: Better help message for cephfs-journal-tool -help command for --rank option.
* Bug #61764: qa: test_join_fs_vanilla is racy
* Bug #61775: cephfs-mirror: mirror daemon does not shutdown (in mirror ha tests)
* Bug #61782: mds: cap revoke and cap update's seqs mismatched
* Bug #61837: qa: test_progress fails with osds full (?)
* Bug #61844: mgr/dashboard: dashboard thread abort
* Bug #61864: mds: replay thread does not update some essential perf counters
* Documentation #61865: add doc on how to expedite MDS recovery with a lot of log segments
* Feature #61866: MDSMonitor: require --yes-i-really-mean-it when failing an MDS with MDS_HEALTH_TRIM or MDS_HEALTH_CACHE_OVERSIZED health warnings
* Bug #61867: mgr/volumes: async threads should periodically check for work
* Bug #61869: pybind/cephfs: holds GIL during rmdir
* Bug #61874: mgr: DaemonServer::ms_handle_authentication acquires daemon locks
* Bug #61879: mds: linkmerge assert check is incorrect in rename codepath
* Bug #61897: qa: rados:mgr fails with MDS_CLIENTS_LAGGY
* Feature #61908: mds: provide configuration for trim rate of the journal
* Bug #61912: mgr hang when purging OSDs
* Bug #61942: The throttle parameter of osd does not take effect for mgr
* Bug #61947: mds: enforce a limit on the size of a session in the sessionmap
* Bug #61950: mds/OpenFileTable: match MAX_ITEMS_PER_OBJ does not honor osd_deep_scrub_large_omap_object_key_threshold
* Bug #61957: test_client_limits.TestClientLimits.test_client_release_bug fails
* Bug #61958: mds: add debug logs for handling setxattr for ceph.dir.subvolume
* Bug #61967: mds: "SimpleLock.h: 417: FAILED ceph_assert(state == LOCK_XLOCK || state == LOCK_XLOCKDONE || state == LOCK_XLOCKSNAP || state == LOCK_LOCK_XLOCK || state == LOCK_LOCK || is_locallock())"
* Bug #61972: cephfs/tools: cephfs-data-scan "cleanup" operation is not parallelised
* Bug #62002: raw list shouldn't list lvm OSDs
* Bug #62013: Object with null version when using versioning and transition
* Bug #62021: mds: unnecessary second lock on snaplock
* Bug #62036: src/mds/MDCache.cc: 5131: FAILED ceph_assert(isolated_inodes.empty())
* Bug #62047: delete image fail on dashboard while it has some protected snapshots
* Bug #62052: mds: deadlock when getattr changes inode lockset
* Bug #62057: mds: add TrackedOp event for batching getattr/lookup
* Bug #62058: mds: inode snaplock only acquired for open in create codepath
* Bug #62075: New radosgw-admin commands to cleanup leftover OLH index entries and unlinked instance objects
* Bug #62076: reef: Test failure: test_grow_shrink (tasks.cephfs.test_failover.TestMultiFilesystems)
* Bug #62077: mgr/nfs: validate path when modifying cephfs export
* Bug #62081: tasks/fscrypt-common does not finish, timesout
* Bug #62084: task/test_nfs: AttributeError: 'TestNFS' object has no attribute 'run_ceph_cmd'
* Feature #62086: mds: print locks when dumping ops
* Bug #62096: mds: infinite rename recursion on itself
* Bug #62114: mds: adjust cap acquistion throttle defaults
* Bug #62146: qa: adjust fs:upgrade to use centos_8 yaml
* Bug #62160: mds: MDS abort because newly corrupt dentry to be committed
* Bug #62164: qa: "cluster [ERR] MDS abort because newly corrupt dentry to be committed: [dentry #0x1/a [fffffffffffffff6,head] auth (dversion lock) v=13..."
* Bug #62187: iozone: command not found
* Bug #62203: gcc-+12: FTBFS random_shuffle is deprecated: use 'std::shuffle' instead 
* Bug #62208: mds: use MDSRank::abort to ceph_abort so necessary sync is done
* Bug #62217: ceph_fs.h: add separate owner_{u,g}id fields
* Bug #62236: qa: run nfs related tests with fs suite
* Bug #62250: retry metadata cache notifications with INVALIDATE_OBJ
* Bug #62265: cephfs-mirror: use monotonic clocks in cephfs mirror daemon
* Bug #62277: Error: Unable to find a match: python2 with fscrypt tests
* Bug #62278: pybind/mgr/volumes: pending_subvolume_deletions count is always zero in fs volume info output
* Bug #62320: lvm list should filter also on vg name
* Bug #62326: pybind/mgr/cephadm: stop disabling fsmap sanity checks during upgrade
* Bug #62355: cephfs-mirror: do not run concurrent C_RestartMirroring context
* Bug #62357: tools/cephfs_mirror: only perform actions if init succeed
* Bug #62382: mon/MonClient: ms_handle_fast_authentication return value ignored
* Bug #62385: mgr/status: wrong kb_used and kb_avail units in command `ceph osd status` results
* Bug #62401: ObjectStore/StoreTestSpecificAUSize.SpilloverFixedTest/2 fails from l_bluefs_slow_used_bytes not matching the expected value
* Bug #62428: cmake: rebuild picks up newer python when originally built with older one
* Feature #62474: rgw: add versioning status during `radosgw-admin bucket stats`
* Bug #62482: qa: "cluster [WRN] Health check failed: 1 pool(s) do not have an application enabled (POOL_APP_NOT_ENABLED)"
* Bug #62492: libcephsqlite: short reads fill 0s at beginning of buffer
* Bug #62494: Lack of consistency in time format
* Bug #62501: pacific(?): qa: mgr-osd-full causes OSD aborts due to ENOSPC (fs/full/subvolume_snapshot_rm.sh)
* Bug #62508: qa: Health check failed: 1 pool(s) do not have an application enabled (POOL_APP_NOT_ENABLED)" in cluster log
* Bug #62510: snaptest-git-ceph.sh failure with fs/thrash
* Bug #62537: cephfs scrub command will crash the standby-replay MDSs
* Bug #62567: postgres workunit times out - MDS_SLOW_REQUEST in logs
* Bug #62577: mds: log a message when exiting due to asok "exit" command
* Bug #62578: mon: osd pg-upmap-items command causes PG_DEGRADED warnings
* Bug #62579: client: evicted warning because client completes unmount before thrashed MDS comes back
* Bug #62580: testing: Test failure: test_snapshot_remove (tasks.cephfs.test_strays.TestStrays)
* Bug #62626: mgr/nfs: include pseudo in JSON output when nfs export apply -i fails
* Bug #62627: global: core fatal signal handler uses may signal-unsafe functions
* Bug #62641: mgr/(object_format && nfs/export): enhance nfs export update failure response
* Bug #62658: error during scrub thrashing: reached maximum tries (31) after waiting for 900 seconds
* Bug #62659: mgr/nfs: report actual errno instead of EIO for single export update failure
* Bug #62663: MDS: inode nlink value is -1 causing MDS to continuously crash
* Bug #62681: high virtual memory consumption when dealing with Chunked Upload
* Bug #62682: mon: no mdsmap broadcast after "fs set joinable" is set to true
* Bug #62700: postgres workunit failed with "PQputline failed"
* Bug #62702: MDS slow requests for the internal 'rename' requests
* Bug #62737: RadosGW API: incorrect bucket quota in response to HEAD /{bucket}/?usage
* Bug #62739: cephfs-shell: remove distutils Version classes because they're deprecated
* Bug #62763: qa: use stdin-killer for ceph-fuse mounts
* Bug #62778: cmake: BuildFIO.cmake should not introduce -std=gnu++17
* Bug #62793: client: setfattr -x ceph.dir.pin: No such attribute
* Bug #62832: common: config_proxy deadlock during shutdown (and possibly other times)
* Bug #62848: qa: fail_fs upgrade scenario hanging
* Bug #62861: mds: _submit_entry ELid(0) crashed the MDS
* Bug #62863: Slowness or deadlock in ceph-fuse causes teuthology job to hang and fail
* Bug #62870: test_nfs task fails due to no orch backend set
* Bug #62873: qa: FAIL: test_client_blocklisted_oldest_tid (tasks.cephfs.test_client_limits.TestClientLimits)
* Bug #62875: SignatureDoesNotMatch when extra headers start with 'x-amzn'
* Feature #62882: mds: create an admin socket command for raising a signal
* Feature #62884: audit: create audit module which persists in RADOS important operations performed on the cluster
* Feature #62892: mgr/snap_schedule: restore scheduling for subvols and groups
* Bug #62925: cephfs-journal-tool: Add preventive measures in the tool to avoid corruting a ceph file system
* Bug #62936: Test failure: test_mirroring_init_failure_with_recovery (tasks.cephfs.test_mirroring.TestMirroring)
* Bug #62938: RGW s3website API prefetches data for range requests
* Bug #62953: qa: fs:upgrade needs updated to upgrade only from N-2, N-1 releases (i.e. reef/quincy)
* Bug #62962: mds: standby-replay daemon crashes on replay
* Bug #62968: mgr/volumes: fix `subvolume group rm` command error message
* Bug #62979: client: queue a delay cap flushing if there are ditry caps/snapcaps
* Bug #63004: CVE-2023-43040 - Improperly verified POST keys.
* Bug #63088: mgr/dashboard: Graphs in Grafana Dashboard are not showing consistent line graphs after upgrading from RHCS 4 to 5.
* Bug #63093: mds: `dump dir` command should indicate that a dir is not cached
* Bug #63103: mds: disable delegating inode ranges to clients
* Bug #63132: qa: subvolume_snapshot_rm.sh stalls when waiting for OSD_FULL warning
* Bug #63141: qa/cephfs: test_idem_unaffected_root_squash fails
* Bug #63154: fs rename must require FS to be offline and refuse_client_session to be set
* Bug #63166: mon/MDSMonitor: metadata not loaded from PAXOS on update
* Bug #63170: rgw: indexless buckets can crash radosgw during sync threads
* Bug #63176: qa: make rank_asok() capable of handling errors from asok commands
* Bug #63188: client: crash during upgrade from octopus to quincy (or from pacific to reef)
* Bug #63195: mgr: remove out&down osd from mgr daemons to avoid warnings
* Bug #63196: compilation fails from git main on ppc64le with missing symbols
* Bug #63218: cmake: dependency ordering error for liburing and librocksdb
* Feature #63224: [RFE] Add an alert for swap space usage
* Bug #63245: rgw/s3select: crashes in test_progress_expressions in run_s3select_on_csv
* Bug #63259: mds: failed to store backtrace and force file system read-only
* Bug #63281: src/mds/MDLog.h: 100: FAILED ceph_assert(!segments.empty())
* Bug #63287: mgr/dashboard: Unable to set max objects under user quota for an user
* Bug #63288: MClientRequest: properly handle ceph_mds_request_head_legacy for ext_num_retry, ext_num_fwd, owner_uid, owner_gid
* Cleanup #63294: mgr: enable per-subinterpreter GIL (Python >= 3.12) 
* Bug #63301: cephfs-data-scan: may remove newest primary link of inode
* Feature #63344: Set and manage nvmeof gw - controller ids ranges
* Bug #63353: resharding RocksDB after upgrade to Pacific breaks OSDs
* Documentation #63354: Add all mon_cluster_log_* configs to docs
* Bug #63360: rgw: rgw-restore-bucket-index does not restore some versioned instance entries
* Bug #63364: MDS_CLIENT_OLDEST_TID: 15 clients failing to advance oldest client/flush tid
* Bug #63388: mgr: discovery service (port 8765) fails if ms_bind ipv6 only
* Bug #63411: qa: flush journal may cause timeouts of `scrub status`
* Bug #63425: tasks.cephadm: ceph.log No such file or directory
* Bug #63433: devicehealth: sqlite3.IntegrityError: UNIQUE constraint failed: DeviceHealthMetrics.time, DeviceHealthMetrics.devid
* Bug #63436: Typo in reshard example
* Feature #63468: mds/purgequeue: add l_pq_executed_ops counter
* Bug #63469: mgr/dashboard: fix rgw multi-site import form helper
* Bug #63482: qa: fs/nfs suite needs debug mds/client
* Bug #63488: smoke test fails from "NameError: name 'DEBUGFS_META_DIR' is not defined"
* Bug #63514: mds: avoid sending inode/stray counters as part of health warning for standby-replay
* Bug #63516: mds may try new batch head that is killed
* Bug #63520: the usage of osd_pg_stat_report_interval_max is not uniform
* Bug #63538: mds: src/mds/Locker.cc: 2357: FAILED ceph_assert(!cap->is_new())
* Bug #63541: Observing client.admin crash in thread_name 'rados' on executing 'rados clearomap..'
* Bug #63557: NVMe-oF gateway prometheus endpoints
* Bug #63561: cephadm: build time install of dependencies fails on build systems that disable network
* Feature #63574: support setting quota in the format of {K|M}iB along with the K|M, {K|M}i
* Bug #63587: Test failure: test_filesystem_sync_stuck_for_around_5s (tasks.cephfs.test_misc.TestMisc)
* Bug #63608: mgr/dashboard: cephfs rename only works when fs is offline
* Bug #63609: osd acquire map_cache_lock high latency
* Bug #63614: cephfs-mirror: the peer list/snapshot mirror status always displays only one mon host instead of all
* Bug #63615: mgr: consider raising priority of MMgrBeacon
* Bug #63619: client: check for negative value of iovcnt before passing it to internal functions during async I/O
* Bug #63629: client: handle context completion during async I/O call when the client is not mounting
* Bug #63632: client: fh obtained using O_PATH can stall the caller during async I/O
* Bug #63633: client: fix copying bufferlist to iovec structures in Client::_read
* Bug #63646: mds: incorrectly issued the Fc caps in LOCK_EXCL_XSYN state for filelock
* Bug #63648: client: ensure callback is finished if write fails during async I/O
* Bug #63658: OSD trim_maps - possible too slow lead to using too much storage space
* Feature #63667: client,libcephfs,cephfs.pyx: add quiesce API
* Bug #63672: qa: nothing provides lua-devel needed by ceph-2:18.0.0-7530.g67eb1cb4.el8.x86_64
* Bug #63679: client: handle zero byte sync/async write cases
* Bug #63680: qa/cephfs: improvements for name generators in test_volumes.py
* Bug #63685: mds: FAILED ceph_assert(_head.empty())
* Bug #63699: qa: failed cephfs-shell test_reading_conf
* Bug #63710: client.5394 isn't responding to mclientcaps(revoke), ino 0x10000000001 pending pAsLsXs issued pAsLsXsFs, sent 30723.964282 seconds ago 
* Bug #63713: mds: encode `bal_rank_mask` with a higher version
* Bug #63722: cephfs/fuse: renameat2 with flags has wrong semantics
* Bug #63727: LogClient: do not output meaningless logs by default
* Bug #63734: client: handle callback when async io fails
* Bug #63740: rgwlc: lock_lambda overwrites ret val
* Bug #63765: change priority of mds rss perf counter to useful
* Bug #63836: OSD:oldest_map/newest_map should be displayed directly in the ceph daemon osd.x status command
* Bug #63882: pybind/mgr/devicehealth: "rados.ObjectNotFound: [errno 2] RADOS object not found (Failed to operate read op for oid $dev"
* Bug #63907: cephfs-mirror: Mirror::update_fs_mirrors crashes while taking lock
* Feature #63945: cephfs_mirror: add perf counters (w/ label) support
* Bug #63952: Boost download URL broken, affecting Windows build
* Documentation #63991: cephfs-top: document all metric fields
* Bug #64012: qa: Command failed qa/workunits/fs/full/subvolume_clone.sh
* Bug #64020: cephadm is not accounting for the memory required nvme gateways are used
* Bug #64042: mgr/snap_schedule: Adding retention which already exists gives improper error message
* Bug #64058: qa: Command failed (workunit test fs/snaps/untar_snap_rm.sh)
* Bug #64061: mds: check the layout in Server::handle_client_mknod
* Bug #64085: qa: stop testing on rhel8
* Bug #64095: ceph-exporter is not included in the deb packages
* Bug #64113: ceph fails to build with Python 3.13: error: there are no arguments to ‘PyEval_CallMethod’ 
* Bug #64124: diff users in the tenant, create the same name topic, rgw topic is covered 
* Bug #64127: mds: passing multiple caps to "fs authorize" cmd causes MON to crash
* Bug #64172: Test failure: test_multiple_path_r (tasks.cephfs.test_admin.TestFsAuthorize)
* Bug #64174: fs/cephadm/renamevolume: volume rename failure
* Bug #64179: client: check for caps issued before incrementing cap ref
* Bug #64191: neorados: compiler warnings in main
* Bug #64209: snaptest-multiple-capsnaps.sh fails with "got remote process result: 1"
* Bug #64236: mon: health store size growing infinitely
* Bug #64290: mds: erroneous "MDS abort because newly corrupt dentry to be committed" because snapclient is not yet synced with snapserver
* Bug #64313: client: do not proceed with I/O if filehandle is invalid
* Bug #64319: OSD does not move itself to crush_location on start, root=default is not applied
* Feature #64322: rgw: enhance radoslist to allow object & object version to be specified
* Feature #64334: The nvmeof gateway has an embedded prometheus exporter than should be scraped by Ceph's Prometheus instance
* Feature #64335: Add alerts to ceph monitoring stack for the nvmeof gateways
* Feature #64387: mds: add per-client perf counters (w/ label) support
* Bug #64440: mds: reversed encoding of MDSMap max_xattr_size/bal_rank_mask v18.2.1 <-> main
* Bug #64456: Missing entries for hardware alerts from the MIB file
* Bug #64479: Memory leak detected when accessing a CephFS volume from Samba using libcephfs
* Bug #64482: ceph: stderr Error: OCI runtime error: crun: bpf create ``: Function not implemented
* Backport #64501: squid: multisite: Deadlock in RGWDeleteMultiObj with default rgw_multi_obj_del_max_aio > 1
* Bug #64548: ceph-base: /var/lib/ceph/crash/posted not chowned to ceph:ceph causing ceph-crash to fail
* Feature #64578: Add a top tool to the nvmeof CLI to support troubleshooting 
* Backport #64601: squid: unittest_rgw_dmclock_scheduler fails for arm64
* Backport #64663: squid: crimson: unittest-seatar-socket failing intermittently
* Bug #64719: SSL session id reuse speedup mechanism of the SSL_CTX_set_session_id_context is not working