# v16.2.7



* Bug #48752: [rfe] generalize ops log output channel for gelf and/or syslog targets
* Feature #50336: mgr/dashboard: "Create Cluster" workflow
* Backport #50483: pacific: segv in AsyncConnection::_stop()
* Backport #50576: pacific: [pwl rwl] IO hang after a period of time I/O of different block sizes
* Backport #50609: pacific: [pwl ssd] indefinite I/O hang if cache is filled to capacity
* Backport #50615: pacific: [pwl ssd] crash when reading a split log entry (after the log wraps)
* Backport #50856: pacific: [pwl ssd] multiple crash / power failure recovery issues
* Backport #51045: pacific: S3 event message eTag misspelled
* Backport #51117: pacific: osd: Run osd bench test to override default max osd capacity for mclock.
* Backport #51274: pacific: mgr/dashboard: deprecated variable usage in grafana dashboards
* Backport #51350: pacific: notifications: notifications stop working after bucket reshard
* Backport #51511: pacific: notification: topic creation is failing when using v4 authentication
* Backport #51597: pacific: valgrind use-after-free in RGWCopyObj::execute() after RGWObjectCtx::invalidate()
* Backport #51646: pacific: osd/OSD: mkfs need wait for transcation completely finish
* Backport #51669: pacific: [pwl ssd] segfault in AbstractWriteLog::get_context() during fio workload
* Backport #51670: pacific: [pwl ssd] first_free_entry corruption on media (segfault in bufferlist::splice() during fio workload)
* Backport #51763: pacific: Missed shared block repair doesn't fix related issues
* Backport #51803: pacific: notifications: zero size in case of delete marker creation
* Backport #51839: pacific: Upgrading the ceph-selinux package unconditionally restarts all running daemons if selinux is enabled
* Backport #51909: pacific: upmap verify failed with pool size decreased
* Backport #51952: pacific: FAILED ceph_assert(attrs || !recovery_state.get_pg_log().get_missing().is_missing(soid) || (it_objects != recovery_state.get_pg_log().get_log().objects.end() && it_objects->second->op == pg_log_entry_t::LOST_REVERT))
* Backport #52021: pacific: mgr/dashboard: add visual regression testing
* Backport #52092: pacific: [pwl] segment fault on syncpoint stack
* Backport #52111: pacific: [pwl] fix buf_persist and add writeback_lat perf counters
* Backport #52251: pacific: cephfs MDS Workload should use rate for counter type metric 
* Backport #52281: pacific: [pwl ssd] pin log entries when reading from disk to avoid a race with retiring
* Backport #52289: pacific: TestModuleSelftest.test_diskprediction_local
* Backport #52292: pacific: mgr/dashboard: False positives in visual regression tests of dashboard component
* Backport #52310: pacific: [pwl ssd] head / tail pointer corruption
* Backport #52322: pacific: LibRadosTwoPoolsPP.ManifestSnapRefcount failure
* Backport #52332: pacific: cmake: ceph-volume build venv is not being created
* Backport #52350: pacific: rgw fix sts memory leak
* Backport #52412: pacific: mgr/snap-schedule: retention set calculation for multiple retention specs is wrong
* Backport #52422: pacific: [pwl ssd] incorrect first_valid_entry calculation in retire_entries()
* Backport #52441: pacific: mds: slow performance on parallel rm operations for multiple kclients
* Backport #52444: pacific: cephfs-mirror: terminating a mirror daemon can cause a crash at times
* Backport #52450: pacific: smart query on monitors
* Backport #52468: pacific: "ERROR: s3tests.functional.test_s3_website.test_website_xredirect_private_relative" in rgw
* Backport #52472: pacific: *: s/virtualenv/python -m venv/
* Bug #52473: mgr/dashboard: added accidently wrong value in api and password string for gaetways
* Backport #52528: pacific: mgr/prometheus: offer ability to disable cache
* Backport #52536: pacific: [pwl ssd] assert in AbstractWriteLog::handle_flushed_sync_point()
* Backport #52559: pacific: pool validation lockup on selfmanaged_snap_create()
* Backport #52569: pacific: [pwl] m_bytes_allocated is calculated incorrectly on reopen
* Backport #52570: pacific: [pwl ssd] memory corruption (shared_ptr related?)
* Backport #52585: pacific: [pwl ssd] m_blocks_to_log_entires not cleared on retiring
* Backport #52588: pacific: lvm activate --all uses systemctl although --no-systemd option is set
* Backport #52598: pacific: Supports specifying osd_id that is not exist and is not in destroyed.
* Backport #52610: pacific: cmake: handle empty CMAKE_THREAD_LIBS_INIT
* Backport #52617: pacific: mgr/dashboard: Incorrect MTU mismatch warning
* Backport #52620: pacific: partial recovery become whole object recovery after restart osd
* Backport #52627: pacific: cephfs-mirror: cephfs-mirror daemon status for a particular FS is not showing
* Backport #52628: pacific: pybind/mgr/volumes: first subvolume permissions set perms on /volumes and /volumes/group
* Backport #52633: pacific: mds,client: add flag to MClientSession for reject reason
* Backport #52639: pacific: MDSMonitor: handle damaged state from standby-replay
* Bug #52667: mgr/dashboard: unable to really completely silence alerts in dashboard
* Backport #52678: pacific: qa: test_kill_mdstable (tasks.cephfs.test_snapshots.TestSnapshots)
* Backport #52679: pacific: "cluster [WRN] 1 slow requests" in smoke pacific
* Backport #52685: pacific: mgr/dashboard: Configuration e2e test failure because of newly added config
* Bug #52697: segmentation faults in libradosgw
* Backport #52733: pacific: [rbd-mirror] unbreak one-way snapshot-based mirroring
* Backport #52743: pacific: mgr/dashboard: make modified API endpoints backward compatible
* Backport #52763: pacific: mgr/dashboard: move force maintenance check to the workflow tests
* Backport #52768: pacific: bluestore repair might cause invalid write
* Backport #52770: pacific: pg scrub stat mismatch with special objects that have hash 'ffffffff'
* Backport #52772: pacific: mgr/dashboard: 'Test failure: test_pwd_expiration_date_update (tasks.mgr.dashboard.test_user.UserTest)'
* Backport #52783: pacific: Cannot perform server-side copy using STS credentials
* Backport #52794: pacific: Progress module optimizations
* Backport #52803: pacific: mgr/dashboard: dashboard 16.2.5 unable to ipv6 wildcard bind
* Backport #52806: pacific: mgr/dashboard: evaluate upgrade to Angular 11
* Bug #52811: mgr/dashboard: mgr crashes when viewing unavailable filesystem info via dashboard
* Fix #52818: Make RGW transaction IDs less deterministic
* Backport #52831: pacific: osd: pg may get stuck in backfill_toofull after backfill is interrupted due to target osd temporarily down
* Backport #52835: pacific: Octopus: tasks.ceph_test_case.TestTimeoutError: Timed out after 60 seconds waiting for HEALTH_OK (currently HEALTH_WARN)
* Backport #52843: pacific: msg/async/ProtocalV2: recv_stamp of a message is set to a wrong value
* Backport #52844: pacific: ceph pg dump times out due to no active mgr 
* Backport #52848: pacific: mgr: mgr crashes when running "rbd perf image iostat" with prometheus rbd metrics enabled
* Backport #52868: stretch mode: allow users to change the tiebreaker monitor
* Backport #52870: pacific: mgr/dashboard: clean up controllers
* Bug #52876: pacific: cluster [WRN] evicting unresponsive client smithi121 (9126), after 303.909 seconds
* Backport #52882: pacific: [pwl] C_WriteSameRequest number_log_entries is not reset from 0
* Backport #52911: pacific: mgr/dashboard: replace "Ceph-cluster" Client connections with active-stdby mgrs
* Backport #52931: pacific: mgr/dashboard: orchestrator/01-hosts.e2e-spec.ts fails
* Backport #52934: pacific: os/bluestore: _do_write_small fix head_pad
* Backport #52936: pacific: Primary OSD crash caused corrupted object and further crashes during backfill after successful write
* Bug #52946: mgr/dashboard: Grafana "Highest Read/Write Latencies" tables show NaN only
* Backport #52960: pacific: Make RGW transaction IDs less deterministic
* Backport #52965: pacific: mgr/dashboard: monitoring: complete grafonnet refactoring
* Backport #52998: pacific: Monitor might crash after upgrade from ceph to 16.2.6
* Backport #52999: pacific: Ceph monitor crash after upgrade from ceph 15.2.14 to 16.2.6
* Backport #53005: pacific: librgw_op_tp.so and librgw_rados_tp.so are included in librgw2 package, but have a different major version number than librgw
* Backport #53006: pacific: RuntimeError: The following counters failed to be set on mds daemons: {'mds.dir_split'}
* Bug #53020: mgr/dashboard: error when closing a form inside the form after angular 11
* Bug #53021: mgr/dashboard: access denied page when a user with read-only permission logs in
* Backport #53026: pacific: mgr/dashboard: make check failure because of conflicting versions
* Backport #53028: pacific: rbd diff between two snapshots lists entire image content with 'whole-object' switch
* Backport #53032: pacific: rbd-mirror: metadata of mirrored image are not properly cleaned up after image deletion
* Bug #53038: mgr/dashboard: cluster expansion review page total memory undefined
* Backport #53065: pacific: mgr/dashboard: wrong data with bonding ethernet in Netork load | Top 10 hosts
* Feature #53077: mgr/dashboard: edit services
* Backport #53091: pacific: [rfe] generalize ops log output channel for gelf and/or syslog targets
* Backport #53092: pacific: mgr/dashboard: running install-deps.sh fails with an error
* Backport #53114: pacific: [pwl ssd] flush cause io re-oreder to writeback layer
* Backport #53118: pacific: [pwl rwl] dead lock issue when pwl initialization failed
* Backport #53120: pacific: client: do not defer releasing caps when revoking
* Backport #53121: pacific: mds: collect I/O sizes from client for cephfs-top
* Backport #53124: pacific: OMAP upgrade to PER-PG format result in ill-formatted OMAP keys.
* Backport #53141: pacific: [pwl] flush requests are dispatched in advance
* Backport #53153: pacific: mgr/dashboard: weird data in monitoring > Alerts details
* Backport #53164: pacific: mds: tcmalloc::allocate_full_cpp_throw_oom(unsigned long)+0xf3)
* Backport #53170: pacific: [pwl] deadlock on AbstractWriteLog::m_lock during shutdown
* Backport #53188: pacific: lvm migrate fails when target is not an lvm device
* Backport #53217: pacific: test: Implement cephfs-mirror trasher test for HA active/active
* Backport #53221: pacific: mgr/dashboard: include mfa_ids in rgw user-details section
* Backport #53231: pacific: MDSMonitor: assertion during upgrade to v16.2.5+
* Backport #53232: pacific: MDSMonitor: no active MDS after cluster deployment
* Backport #53239: pacific: mon: stretch mode blocks kernel clients from connecting
* Backport #53245: pacific: doc: ephemeral pinning with subvolumegroups
* Backport #53253: pacific: mgr/dashboard: Spelling mistake in host-form Network address field 
* Backport #53259: pacific: mon: should always display disallowed leaders when set
* Bug #53263: How CephFS-mirroring work between two region separated clusters ?
* Backport #53264: pacific: [pwl ssd] cache larger than 4G will corrupt itself
* Backport #53270: pacific: mgr/dashboard: edit services
* Backport #53278: pacific: ceph-volume mis-calculates db/wal slot size for clusters that have multiple PVs in a VG
* Bug #53282: mgr/dashboard: daemon events html code minor fixup
* Backport #53284: pacific: get_first_lv() can return wrong lv
* Backport #53291: pacific: rgw: investigate conditional governing cleaning of incomplete multipart uploads
* Backport #53292: pacific: OSD might wrongly attempt to use "slow" device when single device is backing the store
* Backport #53297: pacific: mgr/dashboard: Device health status is not getting listed under hosts section
* Backport #53307: pacific: OMAP upgrade to PER-PG format result in skipped first key.
* Bug #53322: mgr/dashboard: The API toolbar menu entry points to the wrong URL
* Backport #53343: pacific: mgr/dashboard: rgw daemon list: add realm column
* Backport #53347: pacific: qa: v16.2.4 mds crash caused by centos stream kernel
* Backport #53348: pacific: change pg_autoscale profile to scale-up by default
* Bug #53353: mgr/dashboard: orchestrator/03-inventory.e2e-spec.ts failure
* Backport #53363: pacific: mgr/dashboard: daemon events html code minor fixup
* Backport #53364: pacific: mgr/dashboard: provisioned values is misleading in RBD image table
* Backport #53366: pacific: mgr/dashboard: orchestrator/03-inventory.e2e-spec.ts failure
* Backport #53377: pacific: mgr/dashboard: Create host form retain the previously created labels and predefine some common labels
* Bug #53472: Active OSD processes do not see reduced memory target when adding more OSDs
* Bug #53500: rte_eal_init fail will waiting forever
* Bug #53586: rocksdb: build error with rocksdb-6.25.x
* Bug #53589: ceph-volume unstable provisioning behaviour
* Bug #53590: ceph abort at bluefs enospc
* Bug #53671: rgw: can't create a bucket for a user after its S3 credentials deletion/adding
* Bug #53740: mon: all mon daemon always crash after rm pool 
* Bug #53847: `ceph cephadm osd activate` don't work
* Documentation #53871: Can't pull the Ingress daemon due to docker.io rate limit
* Bug #53899: bluefs _allocate allocation failed - BlueFS.cc: 2768: ceph_abort_msg("bluefs enospc")