# v17.2.4



* Backport #55467: quincy: mgr/dashboard:  Grafana e2e tests
* Backport #55552: quincy: rpm: openSUSE build fails - needs explicit gcc version, also can't currently use parquet
* Backport #55653: quincy: mgr/dashboard: rewrite long promql queries using multiline
* Backport #55845: quincy: [rbd-mirror] primary snapshot in-use by replayer can be unlinked and removed
* Backport #55848: quincy: [rbd-mirror] don't prune non-primary snapshot when restarting delta sync
* Backport #55910: quincy: [test] cache-mode/rwl workloads/qemu_xfstests job always times out
* Backport #55936: quincy: client: Inode::hold_caps_until should be a time from a monotonic clock
* Backport #55949: quincy: cephadm crashes when trying to restart an invalid service name which starts with keepalived or haproxy
* Backport #55951: quincy: cephadm: cephadm user/home removed during RPM upgrade
* Backport #55992: quincy: Allow setting crush_device_class in OSD service specs
* Backport #56001: quincy: need to disable -Werror when building PMDK
* Backport #56013: quincy: quota support for subvolumegroup
* Backport #56096: quincy: mgr/DaemonServer.cc: FAILED ceph_assert(pending_service_map.epoch > service_map.epoch)
* Backport #56103: quincy: mgr/volumes: subvolume ls with groupname as '_nogroup' crashes
* Backport #56108: quincy: mgr/volumes: Remove incorrect 'size' in the output of 'snapshot info' command
* Backport #56126: quincy: mgr/dashboard: Dashboard should display some helpful (error) message when the iframe-embedded Grafana dashboard failed to load
* Backport #56142: quincy: [rbd_support] a schedule may get lost due to load vs add race
* Backport #56146: quincy: rbd mirror snapshot schedule is not working properly after a few failover-relocate operations
* Backport #56167: quincy: [rbd-mirror] "failed to unlink local peer from remote image" due to EROFS error
* Backport #56430: quincy: mirror snapshot syncing progress reporting (last_copied_object_number) is broken for sparse images
* Backport #56451: quincy: mgr/dashboard: Update docs for centralized logging
* Backport #56459: quincy: rbd-fuse: librados will filter out -r option from command-line
* Backport #56545: quincy: mgr/dashboard: ingress backend service should list all supported services
* Backport #56563: quincy: mgr/dashboard: rbd paginate image list
* Backport #56569: quincy: [rbd-mirror] bogus "incomplete local non-primary snapshot" replayer error
* Backport #56583: quincy: [pwl][test] make recovery.yaml actually trigger recovery
* Backport #56587: quincy: mgr/dashboard: update cypress to 9.7
* Backport #56593: quincy: mgr/dashboard: alert redirect from passive to active mgr instance doesn't work properly
* Backport #56596: quincy: rbd perf image iostat/iotop lost the ability to gather data across pools
* Backport #56598: quincy: octopus : bluestore_cache_other pool memory leak ?
* Backport #56608: quincy: [test] qemu_dynamic_features_no_cache appears to get stuck
* Backport #56619: quincy: ImageWatcher race condition
* Backport #56641: quincy: Log at 1 when Throttle::get_or_fail() fails
* Backport #56668: quincy: Deferred writes might cause "rocksdb: Corruption: Bad table magic number"
* Backport #56679: quincy: [rbd-mirror] misleading "image is still primary within a remote cluster" error message
* Backport #56681: quincy: disable system_pmdk on s390x
* Backport #56723: quincy: osd thread deadlock
* Backport #56947: quincy: Manager > Object Gateway > Users does not HTML encode "Full Name" field
* Backport #57054: quincy: std::shared_mutex deadlocks on Windows
* Backport #57124: quincy: trash purge schedule subcommands aren't working on el9
* Backport #57184: quincy: crash: pthread_mutex_lock()
* Bug #57191: [cephadm] os tuned profile setting names are not validated from cephadm side
* Bug #57192: [cephadm] os tuned profile placement hosts are not validated
* Backport #57212: quincy: [pwl] AbstractWriteLog::m_lock vs ImageCtx::owner_lock lock ordering
* Backport #57257: quincy: Assert in Ceph messenger
* Bug #57305: bootstrap mgr timeout is too short
* Backport #57352: quincy: use actual monitor addresses when creating a peer bootstrap token
* Backport #57354: quincy: [rbd-mirror] resume pending snapshot replayer shut down when an error is encountered
* Backport #57402: quincy: Incorrect err pointer casts on Windows
* Backport #57436: quincy: mgr/dashboard: Squash is not mandatory field in "Create NFS export" page
* Bug #57442: how detele objects work
* Backport #57453: quincy: RefreshRequest can fail with ENOENT if raced with "rbd flatten"
* Backport #57478: quincy: TestMockMigrationHttpClient and HTTPManager "bind: Address already in use [system:98]" failures
* Backport #57484: quincy: mgr/dashboard: 500 internal server error seen on ingress service creation with unmanaged selected 
* Backport #57684: quincy: undefined reference to "__atomic_load_16" on s390x