# v10.2.8



* Backport #16585: jewel: enormous CLOSE_WAIT connections after re-spawning a mds daemon
* Backport #17768: jewel: transient jerasure unit test failures
* Backport #18193: jewel: transient jerasure unit test failures
* Backport #18321: jewel: librbd::ResizeRequest: failed to update image header: (16) Device or resource busy
* Backport #18496: jewel: Possible deadlock performing a synchronous API action while refresh in-progress
* Backport #18626: jewel: TempURL verification broken for URI encoded object names
* Backport #18669: jewel: [  FAILED  ] TestLibRBD.ImagePollIO in upgrade:client-upgrade-kraken-distro-basic-smithi
* Backport #18683: jewel: mon: 'osd crush move ...' doesnt work on osds
* Backport #18699: jewel: client: fix the cross-quota rename boundary check conditions
* Backport #18705: jewel: fragment space check can cause replayed request fail
* Backport #18775: jewel: Qemu crash triggered by network issues
* Backport #18778: jewel: rbd --pool=x rename y z does not work
* Backport #18792: jewel: Client message throttles are not changeable without restart
* Backport #18815: jewel: rados tool does not work with 0 length input (crashes if '--striper' enabled)
* Backport #18823: jewel: run-rbd-unit-tests.sh assert in lockdep_will_lock, TestLibRBD.ObjectMapConsistentSnap
* Backport #18866: jewel: 'radosgw-admin sync status' on master zone of non-master zonegroup
* Backport #18893: jewel: Incomplete declaration for ContextWQ in librbd/Journal.h
* Backport #18900: jewel: Test failure: test_open_inode
* Backport #18906: jewel: "osd marked itself down" will not recognised if host runs mon + osd on shutdown/reboot
* Backport #18908: jewel: the swift container acl does not support field ".ref"
* Backport #18911: jewel: rbd-nbd: check /sys/block/nbdX/size to ensure kernel mapped correctly
* Backport #18948: jewel: rbd-mirror: additional test stability improvements
* Backport #18949: jewel: mds/StrayManager: avoid reusing deleted inode in StrayManager::_purge_stray_logged
* Backport #18951: jewel: segfault in ceph-osd --flush-journal
* Backport #18957: jewel: ceph-disk: bluestore --setgroup incorrectly set with user
* Backport #18958: jewel: IPv6 Heartbeat packets are not marked with DSCP QoS - simple messenger
* Backport #18969: jewel: Change loglevel to 20 for 'System already converted' message
* Backport #18971: jewel: AdminSocket::bind_and_listen failed after rbd-nbd mapping
* Backport #18972: jewel: ceph-disk does not support cluster names different than 'ceph'
* Backport #18998: jewel: 'ceph auth import -i' overwrites caps, should alert user before overwrite
* Backport #19000: jewel: "osd/PG.cc: 6896: FAILED assert(pg->is_acting(osd_with_shard) || pg->is_up(osd_with_shard))" in rados/upgrade
* Backport #19044: jewel: buffer overflow in test LibCephFS.DirLs
* Backport #19048: jewel: multisite: some yields in RGWMetaSyncShardCR::full_sync() resume in incremental_sync()
* Backport #19062: jewel: Build ceph-resource-agents package for rpm based os
* Backport #19083: jewel: osd: preserve allocation hint attribute during recovery
* Backport #19142: jewel: Ceph Xenial Packages - ceph-base missing dependency for psmisc
* Backport #19145: jewel: rgw: a few cases where rgw_obj is incorrectly initialized
* Backport #19155: jewel: rgw: typo in rgw_admin.cc
* Backport #19158: jewel: RGW health check errors out incorrectly 
* Backport #19163: jewel: radosgw-admin: add the 'object stat' command to usage
* Backport #19171: jewel: rgw: S3 create bucket should not do response in json
* Backport #19183: jewel: os/filestore/HashIndex: be loud about splits
* Backport #19206: jewel: Invalid error code returned by MDS is causing a kernel client WARNING
* Backport #19210: jewel: pre-jewel "osd rm" incrementals are misinterpreted
* Backport #19211: jewel: rgw: "cluster [WRN] bad locator @X on object @X...." in cluster log 
* Backport #19223: jewel: osd crashes during hit_set_trim and hit_set_remove_all if hit set object doesn't exist
* Backport #19311: segmentation fault when call do_fiemap() in filestore
* Backport #19314: jewel: osd: pg log split does not rebuild index for parent or child
* Backport #19321: jewel: multisite: possible infinite loop in RGWFetchAllMetaCR
* Backport #19325: jewel: [api] temporarily restrict (rbd_)mirror_peer_add from adding multiple peers
* Backport #19328: jewel: osd_snap_trim_sleep option does not work
* Backport #19330: jewel: upgrade to multisite v2 fails if there is a zone without zone info
* Backport #19332: jewel: brag fails to count "in" mds
* Backport #19334: jewel: MDS heartbeat timeout during rejoin, when working with large amount of caps/inodes
* Backport #19352: jewel: RadosImport::import should return an error if Rados::connect fails
* Backport #19353: jewel: multisite: some 'radosgw-admin data sync' commands hang
* Backport #19355: jewel: when converting region_map we need to use rgw_zone_root_pool
* Backport #19357: jewel: systemctl stop rbdmap unmaps all rbds and not just the ones in /etc/ceph/rbdmap
* Backport #19358: jewel: rbdmap documentation is inadequate
* Backport #19392: jewel:mon: remove bad rocksdb option
* Backport #19394: OSD blocks all readonly ops when OSD reaches full
* Backport #19404: jewel: core: two instances of omap_digest mismatch
* Backport #19461: jewel: admin ops: fix the quota section
* Backport #19464: jewel: monitor creation with IPv6 public network segfaults
* Backport #19468: jewel: [api] is_exclusive_lock_owner doesn't detect that is has been blacklisted
* Backport #19469: jewel: rgw_file:  leaf objects (which store Unix attrs) can be deleted when children exist
* Backport #19474: jewel: multisite: EPERM when trying to read SLO objects as system/admin user
* Backport #19476: jewel: RGW S3 v4 authentication issue with X-Amz-Expires
* Backport #19478: jewel: "zonegroupmap set" does not work
* Backport #19481: jewel: ceph degraded and misplaced status output inaccurate
* Backport #19482: jewel: No output for "ceph mds rmfailed 0 --yes-i-really-mean-it" command
* Backport #19484: jewel: New added OSD always down when full flag is set
* Backport #19493: jewel: ceph-disk: Racing between partition creation & device node creation
* Backport #19495: jewel: Objecter::epoch_barrier isn't respected in _op_submit()
* Backport #19508: Upgrading from 0.94.6 to 10.2.6 can overload monitors (failed to encode map with expected crc)
* Backport #19523: jewel: `radosgw-admin zone create` command with specified zone-id creates a zone with different id 
* Backport #19525: jewel: rgwfs hung due to missing unlock within unlink operation
* Backport #19536: jewel: ceph-disk list reports mount error for OSD having mount options with SELinux context
* Tasks #19538: jewel v10.2.8
* Backport #19546: jewel integration branch fails to build for centos (regression)
* Backport #19547: rbdmap.service not included in debian packaging (jewel-only)
* Backport #19562: jewel: "api_misc: [  FAILED  ] LibRadosMiscConnectFailure.ConnectFailure"
* Backport #19575: jewel: rgw: unsafe access in RGWListBucket_ObjStore_SWIFT::send_response()
* Backport #19607: jewel: multisite: fetch_remote_obj() gets wrong version when copying from remote
* Backport #19610: jewel: [librados_test_stub] cls_cxx_map_get_XYZ methods don't return correct value
* Backport #19612: jewel: Issues with C API image metadata retrieval functions
* Backport #19617: jewel: common/LogClient.cc: 310: FAILED assert(num_unsent <= log_queue.size())
* Backport #19619: jewel: MDS server crashes due to inconsistent metadata.
* Backport #19646: jewel: ceph-disk: directory-backed OSDs do not start on boot
* Backport #19660: jewel: rgw_file: fix readdir after dir-change
* Backport #19662: jewel: rgw_file: fix event expire check, don't expire directories being read
* Backport #19665: jewel: C_MDSInternalNoop::complete doesn't free itself
* Backport #19666: jewel: fs：The mount point break off when mds switch hanppened.
* Backport #19668: jewel: MDS goes readonly writing backtrace for a file whose data pool has been removed
* Backport #19671: jewel: MDS assert failed when shutting down
* Backport #19673: jewel: cephfs: mds is crushed, after I set about 400 64KB xattr kv pairs to a file
* Backport #19675: jewel: cephfs: Test failure: test_data_isolated (tasks.cephfs.test_volume_client.TestVolumeClient)
* Backport #19677: jewel: Jewel ceph-fuse does not recover after lost connection to MDS
* Backport #19686: jewel: Give requested scrubs a higher priority
* Backport #19690: jewel: Improvements to crushtool manpage
* Backport #19701: jewel: osd/PGLog.cc: 1047: FAILED assert(oi.version == i->first)
* Backport #19709: jewel: Enable MDS to start when session ino info is corrupt
* Backport #19711: jewel: [test] test_notify.py: rbd.InvalidArgument: error updating features for image test_notify_clone2
* Backport #19722: jewel: rgw_file: introduce rgw_lookup type hints
* Backport #19724: jewel: RGW S3 v4 authentication issue with X-Amz-Expires
* Backport #19727: jewel: rbd-nbd: immediate seg fault starting the daemon
* Backport #19728: jewel: rgw: add radosgw-admin command to check progress toward bucket sharding limits
* Backport #19736: radosgw/s3 chunked transfer encodings and fast_forward_request
* Backport #19757: jewel: fix failed to create bucket if a non-master zonegroup has a single zone
* Backport #19762: jewel: non-local cephfs quota changes not visible until some IO is done
* Backport #19772: jewel: rgw: swift: disable revocation thread under certain circumstances
* Backport #19774: jewel: osd: promote throttle parameters are reversed
* Backport #19786: jewel: ceph jewel  failed create s3 tpye subuser from admin rest api
* Backport #19806: jewel: APIs to support Ragweed suite
* Backport #19846: jewel: write to cephfs mount hangs, ceph-fuse and kernel
* Backport #20014: jewel: multisite: bi_list() decode failures
* Backport #20027: jewel: Deadlock on two ceph-fuse clients accessing the same file
* Backport #20032: jewel: osd_scrub_sleep option blocks op thread in jewel + later
* Backport #20078: rgw jewel valgrind failure: make_params
* Backport #20088: valgrind reports leak in RGWRemoteMetadataCR
* Backport #20126: Jewel: Can't repair when only an attr object error
* Backport #20140: jewel: Journaler may execute on_safe contexts prematurely
* Backport #20148: jewel: Too many stat ops when MDS trying to probe a large file
* Backport #20151: jewel: ceph-disk fails if OSD udev rule triggers prior to mount of /var
* Backport #20157: jewel: rgw_file: handle chunked readdir
* Backport #20172: jewel: PR #14886 creates a SafeTimer thread per PG
* Backport #20401: jewel: "MaxWhileTries: 'wait_until_healthy' reached maximum tries (150) after waiting for 900 seconds" in upgrade/hammer-jewel-x/parallel
* Backport #20412: test_remote_update_write (tasks.cephfs.test_quota.TestQuota) fails in Jewel 10.2.8 integration testing
* Bug #20502: crush: Jewel upgrade misbehaving with custom roots/rulesets
* Backport #20578: jewel: mon: fix 'sortbitwise' warning on jewel
* Feature #20733: RGW  bucket limits