# v16.2.5



* Bug #49962: 'sudo ceph --cluster ceph osd crush tunables default' fails due to valgrind: Unknown option: --exit-on-first-error=yes
* Backport #49982: pacific: BlueRocksEnv::GetChildren may pass trailing slashes to BlueFS readdir
* Backport #49992: pacific: unittest_mempool.check_shard_select failed
* Backport #50014: pacific: ubuntu ceph-mgr package does not pickup libsqlite3-mod-ceph automatically
* Backport #50034: pacific: mgr/dashboard: mute i18n output
* Backport #50183: pacific: client: openned inodes counter is inconsistent
* Backport #50186: pacific: qa: daemonwatchdog fails if mounts not defined
* Backport #50241: pacific: cephfs-mirror: update docs with `fs snapshot mirror daemon status` interface
* Backport #50254: pacific: mds: standby-replay only trims cache when it reaches the end of the replay log
* Bug #50258: pacific: qa: "run() got an unexpected keyword argument 'stdin_data'"
* Backport #50282: pacific: MDS slow request lookupino #0x100 on rank 1 block forever on dispatched
* Backport #50289: pacific: MDS stuck at stopping when reducing max_mds
* Backport #50294: pacific: mgr/prometheus: Update ceph_pool_* metrics to include additional labels
* Backport #50301: pacific: rgw: radoslist incomplete multipart parts marker
* Feature #50318: mgr/dashboard: when adding host, allow to specify network address and labels
* Backport #50344: pacific: mon: stretch state is inconsistently-maintained on peons, preventing proper failover in repeated netsplits
* Backport #50358: pacific: npm problem causes "make-dist" to fail when directory contains colon character
* Bug #50369: mgr/volumes/nfs: drop type param during cluster create
* Backport #50392: pacific: cephfs-top: exception: addwstr() returned ERR
* Backport #50406: pacific: mon: new monitors may direct MMonJoin to a peon instead of the leader
* Backport #50427: pacific: Remove erroneous elements in hosts-overview Grafana dashboard
* Backport #50445: pacific: client: Inode.cc: 405: FAILED ceph_assert(_ref >= 0)
* Backport #50488: pacific: mgr/nfs: move nfs code out of volumes plugin
* Backport #50505: pacific: mon/MonClient: reset authenticate_err in _reopen_session()
* Backport #50537: pacific: "ceph fs snapshot mirror daemon status" should not use json keys as value
* Backport #50538: pacific: mgr/pybind/snap_schedule: do not fail when no fs snapshots are available
* Backport #50541: pacific: libcephfs: support file descriptor based *at() APIs
* Backport #50543: pacific: mgr/dashboard: fix HAProxy (now called ingress)
* Bug #50579: pacific: Ceph SIGILL on RaspberryPi 4 aarch64
* Backport #50597: pacific: mgr/nfs: Add troubleshooting section
* Backport #50602: pacific: osd: check_full_status: check don't cares about RocksDB size
* Backport #50606: pacific: osd/scheduler/mClockScheduler: Async reservers are not updated with the overridden values for "osd_max_backfills" 
* Backport #50624: pacific: qa: "ls: cannot access 'lost+found': No such file or directory"
* Backport #50627: pacific: client: access(path, X_OK) on non-executable file as root always succeeds
* Backport #50629: pacific: cephfs-mirror: ignore snapshots on parent directories when synchronizing snapshots
* Backport #50630: pacific: mds: Error ENOSYS: mds.a started profiler
* Backport #50632: pacific: mds: failure replaying journal (EMetaBlob)
* Backport #50636: pacific: session dump includes completed_requests twice, once as an integer and once as a list
* Backport #50662: pacific: mgr/dashboard: remove NFSv3 support from dashboard
* Backport #50664: pacific: mgr/dashboard: Update the embedded grafana dashboard to show compressions stats by pool
* Backport #50666: pacific: upgrade:nautilus-x-pacific: LibRadosService.StatusFormat failure
* Backport #50673: pacific: [test] qa/workunits/rbd: use bionic version of qemu-iotests for focal
* Backport #50674: pacific: mgr/dashboard: fix bucket versioning when locking is enabled
* Backport #50680: pacific: mgr/dashboard: avoid data processing in crush-map component
* Backport #50698: pacific: fix 32-bit/64-bit server/client interoperability under msgr2
* Backport #50702: pacific: Data loss propagation after backfill
* Backport #50706: pacific: _delete_some additional unexpected onode list
* Backport #50710: pacific: rgw: radosgw-admin bucket rm --bucket <bucket> --purge-objects
* Backport #50713: pacific: librbd: removing a snapshot with multiple peer can go into an infinite loop
* Backport #50718: pacific: [pwl] cache can't be opened after a crash or power failure
* Backport #50726: pacific: Boost download url has changed, breaking Windows build
* Backport #50754: pacific: win32*.sh: disable libcephsqlite when targeting Windows
* Backport #50757: pacific: [pwl] "rbd status" output is incorrect
* Backport #50764: pacific: mgr/dashboard: add Services e2e tests
* Backport #50767: pacific: mgr/dashboard: bucket name constraints
* Backport #50769: pacific: mgr/dashboard: RGW buckets async validator slow performance
* Backport #50772: pacific: mgr/dashboard: Physical Device Performance grafana graphs for OSDs do not display
* Backport #50773: pacific: mgr/dashboard: typo: "Filesystems" to "File Systems"
* Backport #50774: pacific: mgr/dashboard: ingress service creation follow-up
* Backport #50779: pacific: mgr/progress: progress can be negative
* Backport #50782: pacific: AvlAllocator.cc: 60: FAILED ceph_assert(size != 0)
* Backport #50791: pacific: osd: write_trunc omitted to clear data digest 
* Backport #50794: pacific: osd: FAILED ceph_assert(recovering.count(*i)) after non-primary osd restart when in backfill_unfound state
* Backport #50797: pacific: mon: spawn loop after mon reinstalled
* Bug #50835: crc32c unittest failed on aarch64 with gcc V8.3 when HAVE_ARMV8_CRC_CRYPTO_INTRINSICS is false
* Backport #50843: pacific: mgr/nfs: cli is broken as cluster id and binding arguments are optional
* Backport #50846: pacific: mds: journal recovery thread is possibly asserting with mds_lock not locked
* Backport #50848: pacific: mds: "cluster [ERR]   Error recovering journal 0x203: (2) No such file or directory" in cluster log"
* Backport #50864: pacific: mgr/dashboard: query params is lost when the login is redirected to the change-password page
* Backport #50871: pacific: cephfs-mirror: use sensible mount/shutdown timeouts
* Backport #50872: pacific: qa: testing kernel patch for client metrics causes mds abort
* Backport #50873: pacific: mon,doc: deprecate min_compat_client
* Backport #50875: pacific: mds: MDSLog::journaler pointer maybe crash with use-after-free
* Backport #50876: pacific: cephfs-mirror: allow mirror daemon to connect to local/primary cluster via monitor address
* Backport #50877: pacific: qa: test_mirroring_init_failure_with_recovery failure
* Backport #50881: pacific: mgr/dashboard: upgrade frontend deps due to security vulnerabilities
* Backport #50882: pacific: mgr/dashboard: fix ESOCKETTIMEDOUT E2E failure
* Backport #50899: pacific: mds: monclient: wait_auth_rotating timed out after 30
* Backport #50902: pacific: mgr/dashboard: provide the service events when showing a service in the UI
* Backport #50913: pacific: MDS heartbeat timed out between during executing MDCache::start_files_to_recover()
* Backport #50917: pacific: Mirroring path "remove" don't not seem to work
* Backport #50929: pacific: mgr/dashboard: 'grafana dashboards update' command fails with 'Dashboard not found'
* Backport #50938: pacific: OSDs broken after nautilus->octopus upgrade: rocksdb Corruption: unknown WriteBatch tag
* Backport #50942: pacific: rbd-mirror: segfault on snapshot replayer on shutdown
* Backport #50959: pacific: mgr/dashboard: drop container image name and id from services list
* Backport #50962: pacific: mgr/dashboard: fix API docs link 
* Backport #50968: pacific: mgr/dashboard: create NFS export > RGW tenanted user id shown without tenant prefix
* Bug #50971: mgr/dashboard: crushmap tree doesn't display crush type other than root
* Backport #50982: pacific: Pacific - RadosGW not starting after upgrade
* Backport #50986: pacific: unaligned access to member variables of crush_work_bucket
* Backport #50989: pacific: mon: slow ops due to osd_failure
* Backport #50993: pacific: cephfs-mirror: incrementally transfer snapshots whenever possible
* Backport #51001: pacific: Reproducible crash on multipart upload to bucket with policy
* Bug #51026: mgr/dashboard: Grafana datasource not filled in
* Backport #51036: pacific: mgr/dashboard: Grafana datasource not filled in
* Backport #51037: pacific: mgr/dashboard: API Version changing doesnt affect to pre-defined methods
* Backport #51038: pacific: mgr/dashboard: dashboard not working with a URL prefix
* Backport #51040: pacific: bluefs _allocate unable to allocate, though enough free
* Backport #51046: pacific: qemu task fails to install packages, workload isn't run
* Backport #51053: pacific: mgr/dashboard: partially deleted RBDs are only listed by CLI
* Backport #51058: pacific: "trash purge" shouldn't stop at the first unremovable image
* Backport #51063: pacific: mgr/dashboard: fix bucket objects and size calculations
* Backport #51070: pacific: qa: test_ephemeral_pin_distribution failure
* Backport #51084: pacific: mgr/volumes: hang when removing subvolume when pools are full
* Backport #51085: pacific: mds: scrub error on inode 0x1
* Backport #51086: pacific: qa: fs:mirror: reduced data availability
* Backport #51087: pacific: mds: segfault printing unknown metric
* Backport #51094: pacific: mgr crash loop after increase pg_num
* Backport #51105: pacific: batch ignores bluestore_block_db_size in ceph.conf
* Backport #51135: pacific: mgr/dashboard: Bucket creation fails when selecting locking with certain values
* Backport #51138: pacific: rgw: completion of multipart upload leaves delete marker
* Backport #51143: pacific: directories with names starting with a non-ascii character disappear after reshard
* Backport #51148: pacific: mgr/dashboard: when adding host, allow to specify network address and labels
* Bug #51170: pacific: qa: AttributeError: 'RemoteProcess' object has no attribute 'split'
* Backport #51180: pacific: mgr/Dashboard: right Navigation should work on click when page width is less than768 px
* Backport #51185: pacific: qa: update RHEL to 8.4
* Backport #51186: pacific: qa: add async dirops testing
* Backport #51203: pacific: mds: CephFS kclient gets stuck when getattr() on a certain file
* Backport #51215: pacific: Global Recovery Event never completes
* Backport #51230: pacific: qa: fs:bugs does not specify distro
* Backport #51231: pacific: pybind/mgr/snap_schedule: Invalid command: Unexpected argument 'fs=cephfs'
* Backport #51232: pacific: qa: scrub code does not join scrubopts with comma
* Backport #51233: pacific: pybind/cmd_argparse: does not process vector of choices
* Backport #51235: pacific: doc: pacific updates
* Backport #51251: pacific: qa: fs:upgrade uses teuthology default distro
* Backport #51259: pacific: mgr/dashboard: 500 server error when exiting out of maintenance mode
* Backport #51273: pacific: mgr/dashboard: bucket details: show lock retention period only in days
* Backport #51283: pacific: cephfs-mirror: disallow adding a active peered file system back to its source
* Backport #51284: pacific: cephfs-mirror: false warning of "keyring not found" seen in cephfs-mirror service status is misleading
* Backport #51286: pacific: MDSMonitor: crash when attempting to mount cephfs
* Bug #51312: mgr/dashboard: Alertmanager fails to POST alerts
* Backport #51313: pacific: osd:scrub skip some pg
* Backport #51322: pacific: qa: test_data_scan.TestDataScan.test_pg_files AssertionError: Items in the second set but not the first:
* Backport #51324: pacific: pacific: client: abort after MDS blocklist
* Backport #51326: pacific: mgr/dashboard: Alertmanager fails to POST alerts
* Backport #51335: pacific: mds: avoid journaling overhead for setxattr("ceph.dir.subvolume") for no-op case
* Backport #51343: pacific: mgr/dashboard: crushmap tree doesn't display crush type other than root
* Backport #51345: pacific: TestMockMigrationHttpClient.OpenResolveFail started to fail because invalid.ceph.com is now resolvable
* Backport #51371: pacific: OSD crash FAILED ceph_assert(!is_scrubbing())
* Backport #51412: pacific: mds: mkdir on ephemerally pinned directory sometimes blocked on journal flush
* Backport #51413: pacific: cephfs-mirror: do not terminate on SIGHUP
* Backport #51414: pacific: mds: "FAILED ceph_assert(r == 0 || r == -2)"
* Backport #51447: pacific: mgr/dashboard: bucket name input takes spaces eventhough its not allowed
* Backport #51453: pacific: Add simultaneous scrubs to rados/thrash
* Backport #51455: pacific: progress module: TypeError: '<' not supported between instances of 'str' and 'int'
* Backport #51475: pacific: mgr/dashboard: User database migration has been cut out
* Backport #51478: pacific: Incorrect OSD out count on landing page
* Backport #51489: pacific: mgr/dashboard: run cephadm orch backend E2E tests
* Bug #51672: mgr/dashboard: test_dashboard_e2e.sh: "AssertionError: Timed out retrying: Expected to find content: 'Continue' within the selector: 'cd-modal button' but never did."