File _patchinfo of Package patchinfo.7200
<patchinfo incident="7200">
<issue id="1024691" tracker="bnc">systemctl stop rbdmap unmaps ALL rbds and not just the ones in /etc/ceph/rbdmap</issue>
<issue id="1029482" tracker="bnc">rbdmap not present in ceph-common with SLES12 SP1</issue>
<issue id="1047977" tracker="bnc">nfs-ganesha.service: Failed with result 'core-dump'.</issue>
<issue id="1015371" tracker="bnc">swift | This returns with HTTP/1.1 401 Unauthorized</issue>
<issue id="1042973" tracker="bnc">ceph systemd dependencies are wrong ("Transaction order is cyclic")</issue>
<issue id="1043399" tracker="bnc">some OSDs are down with assigning the rbd image in EC pool to a VM instance</issue>
<issue id="1047020" tracker="bnc">Remaining packages after upgrade SES4 -> SES5 M7</issue>
<issue id="1050063" tracker="bnc">typos in rgw output line in "ceph status"</issue>
<issue id="1028109" tracker="bnc">Invalid error code returned by MDS is causing a kernel client WARNING</issue>
<issue id="1047244" tracker="bnc">cmpext is unsupported atop EC data pools</issue>
<issue id="997082" tracker="bnc">OSD disk failures take a long time to be propagated to the monitors and clients / gateways</issue>
<issue id="1033786" tracker="bnc">documentation: man crushtool does not cover the "--show-mappings" parameter</issue>
<issue id="970642" tracker="bnc">'ceph-deploy mds destroy' not implemented</issue>
<issue id="1025643" tracker="bnc">OSDs fail to start after server reboot</issue>
<category>recommended</category>
<rating>moderate</rating>
<packager>smithfarm</packager>
<description>This update for ceph fixes the following issues:
* fs: Invalid error code returned by MDS is causing a kernel client WARNING
(issue#19205, pr#13831, boo#1028109)
* rbd: systemctl stop rbdmap unmaps all rbds and not just the ones in
/etc/ceph/rbdmap (boo#1024691)
* doc: Improvements to crushtool manpage (boo#1033786)
* rgw: Use decoded URI when verifying TempURL (boo#1015371)
* tools: ceph-disk: do not setup_statedir on trigger (boo#1025643)
+ Async Messenger patches from PTF boo#997082
* DispatchQueue: Move from msg/simple to msg
* DispatchQueue: move dispatch_throtter from SimpleMessenger to DispatchQueue
* AsyncMessenger: introduce DispatchQueue to separate nonfast message
* AsyncConnection: make local message deliver via DispatchQueue
* AsyncConnection: change all exception deliver to DispatchQueue
* AsyncConnection: add DispathQueue throttle
* DispatchQueue: add name to separte different instance
* DispatchQueue: remove pipe words
* AsyncConnection: release dispatch throttle with fast dispatch message
* AsyncConnection: continue to read when meeting EINTR
* AsyncConnection: need to prepare message when features mismatch
* AsyncConnection: fix delay state using dispatch_queue
* msg/simple: add ms_handle_refused callback
* OSD: Implement ms_handle_refused
* messages/MOSDFailure.h: distinguish between timeout and immediate failure
* msg/async: implement ECONNREFUSED detection
* test/osd: add test for fast mark down functionality
- Note that boo#970642 is fixed by commit fc3e6567416769f83d7b258f8dc852f96f2adeaf
"doc: ceph-deploy man: remove references to mds destroy. Not implemented"
which is included in this update
- Update to version 12.1.2+git.1501865907.88dc8cdc9d:
+ latest upstream master (7350fa65d79b5f0ace427e609565f07be35cfe62)
* os/bluestore: allow multiple DeferredBatches in flight at once
(pr#16769, issue#20295, boo#1043399)
- Update to version 12.1.2+git.1501790037.82d6a9c37a:
+ osd: move CMPEXT response munging into completion function
(boo#1047244)
- Update to version 12.1.2+git.1501684697.aa6b0153ef:
+ upstream luminous v12.1.2 release candidate
+ latest upstream master (e27a78c9ed3984b54a312ced3ad67fcc8f464fc0)
* mon: add mgr metdata commands, and overall 'versions' command for all
daemon versions (boo#1050063, pr#16460)
- Update to version 12.1.1+git.1501231943.67577225bc
+ Notable changes:
* rgw_file: properly & |'d flags (boo#1047977)
- Update to version 12.1.0+git.1500295948.39d2e5f424:
+ spec file:
* socat build dependency only for make check
* fix typo WTIH_BABELTRACE
- Update to version 12.1.0+git.1500284183.1ed97736ab:
+ Spec file changes:
* bump epoch (not relevant for SUSE)
* add socat buildrequires
* libcephfs2 now obsoletes libcephfs1 (boo#1047020)
* HAVE_BABELTRACE option renamed to WITH_BABELTRACE
* fix a regression in ceph-base %postun scriptlet
+ Selected other changes:
* Revert "msg/async: increase worker reference with local listen table enabled backend"
* build: build erasure-code isa lib without versions
* ceph-disk/ceph_disk/main.py: Replace ST_ISBLK() test by is_diskdevice()
* ceph.in: filter out audit from ceph -w
* common,config: OPT_FLOAT and OPT_DOUBLE output format in config show
* common/Timer: do not add event if already shutdown
* compressor/zlib: remove g_ceph_context/g_conf from compressor plugin
* core: make the conversion from wire error to host OS work
* core:" Stringify needs access to << before reference" src/include/stringify.h
* crush/CrushWrapper: make get_immediate_parent[_id] ignore per-class shadow hierarchy
* doc/mgr/dashboard: update dashboard docs to reflect new defaults
* doc/rados.8: add offset option for put command
* doc: fix link for ceph-mgr cephx authorization
* doc: update intro, quick start docs
* isa-l: update isa-l to v2.18
* librbd: fail IO request when exclusive lock cannot be obtained
* mgr,librados: service map
* mgr,mon: debug init and mgrdigest subscriptions
* mgr,mon: enable/disable mgr modules via 'ceph mgr module ...' commands
* mgr/ClusterState: do not mangle PGMap outside of Incremental
* mgr/restful: bind to :: and update docs
* mgr: Zabbix monitoring module
* mon,crush: create crush rules using device classes for replicated and ec pools via cli
* mon/MgrStatMonitor: do not crash on luminous dev version upgrades
* mon/MonClient: respect the priority in SRV RR
* mon/Monitor: recreate mon session if features changed
* mon/OSDMonitor: do not allow crush device classes until luminous
* mon/OSDMonitor: guard 'osd crush set-device-class'
* mon/mgr: add detail error infomation
* mon: Fix status output warning for mon_warn_osd_usage_min_max_delta
* mon: add support public_bind_addr option
* mon: collect mon metdata as part of the election
* mon: debug session feature tracking
* mon: do crushtool test with fork and timeout, but w/o exec of crushtool
* mon: fix hang on deprecated/removed 'pg set_\*full_ratio' commands
* mon: include device class in tree view; hide shadow hierarchy
* mon: maintain the "cluster" PerfCounters when using ceph-mgr
* mon: revamp health check/warning system
* msg/async: increase worker reference with local listen table enabled backend
* msg/async: make recv_stamp more precise
* msg: QueueStrategy::wait() joins all threads
* msg: make listen backlog an option, increase from 128 to 512
* os/blueStore: Failure retry for opening file
* os/bluestore/BlueFS: clean up log_writer aios from compaction
* os/bluestore/BlueFS: clear current log entrys before dump all fnode
* os/bluestore/KernelDevice: batch aio submit
* os/bluestore: Make BitmapFreelistManager kv itereator short lived
* os/bluestore: cap rocksdb cache size
* os/bluestore: clear up redundant size assignment in KernelDevice
* os/bluestore: differ default cache size for hdd/ssd backends
* os/bluestore: fix deferred_aio deadlock
* os/bluestore: move object exist in assign nid
* os/bluestore: only submit deferred if there is any
* os/bluestore: reduce some overhead for _do_clone_range() and _do_remove()
* os/bluestore: use bufferlist functions whenever possible
* os/filestore/FileJournal: FileJournal::open() close journal file before return error
* os/filestore/HashIndex: randomize split threshold by a configurable amount
* os/filestore: call committed_thru when no journal entries are replayed
* os/filestore: do not free event if not added
* os/filestore: queue ondisk completion before apply work
* os/filestore: require experimental flag for btrfs
* osd/OSD: auto class on osd start up
* osd/PrimaryLogPG solve cache tier osd high memory consumption
* osd/ReplicatedBackend: reset thread heartbeat after every omap entry
* osd/filestore: Revert "os/filestore: move ondisk in front
* osd: Check for and automatically repair object info soid during scrub
* osd: Preserve OSDOp information for historic ops
* osd: check queue_transaction return value
* osd: clear_queued_recovery() in on_shutdown()
* osd: compact osd feature
* pybind/mgr/dashboard: bind to :: by default
* pybind/mgr/dashboard: fix get kernel_version error
* rbd-mirror: ignore permission errors on rbd_mirroring object
* rbd: do not attempt to load key if auth is disabled
* rgw: fix compilation with cryptopp
* rgw,test: fix rgw placement rule pool config option
* rgw/rgw_frontend.h: Return negative value for empty uid in RGWLoadGenFrontend::init()
* rgw/rgw_rados: Remove duplicate calls in RGWRados::finalize()
* rgw: /info claims we do support Swift's accounts ACLs
* rgw: Correcting the condition in ceph_assert while parsing an AWS Principal
* rgw: Do not fetch bucket stats by default upon bucket listing
* rgw: Fix a bug that multipart upload may exceed the quota
* rgw: Initialize pointer fields
* rgw: fix leaks with incomplete multiparts
* rgw: fix memory leak in copy_obj_to_remote_dest
* rgw: fix not initialized vars which cause rgw crash with ec data pool
* rgw: fix radosgw-admin data sync run crash
* rgw: fix s3 object uploads with chunked transfers and v4 signatures
* rgw: fix the subdir without slash of s3 website url
* rgw: fix zone did't update realm_id when added to zonegroup
* rgw: log_meta only for more than one zone
* rgw: rgw_file skip policy read for virtual components
* rgw: using RGW_OBJ_NS_MULTIPART in check_bad_index_multipart
* rgw: verify md5 in post obj
* rgw_file: add compression interop to RGW NFS
* rgw：fix s3 aws v2 signature priority between header['X-Amz-Date'] and header['Date']
* vstart.sh: bind restful, dashboard to ::, not 127.0.0.1
* vstart: add --rgw_compression to set rgw compression plugin
* vstart: allow to start multiple radosgw when RGW=x
- Update to version 12.1.0+git.1499090920.2c11b88d14:
+ With this update, Ceph no longer autocreates a pool called "rbd" with id 0
Deployment tools must take this into account!
+ systemd: Add explicit Before=ceph.target (boo#1042973)
- Update to version 12.1.4+git.1502953626.155adb5253:
+ ceph.spec: provide libxio 1.7 and obsolete libxio <= 1.7
- Update to version 12.1.4+git.1502878199.4183317204:
+ ceph.spec: obsolete libxio
(Note to reviewer: libxio was only ever used by Ceph (for an experimental
IB/RDMA feature), it was never submitted to Factory, and its functionality has
been integrated into Ceph in a different form. For all practical purposes,
libxio has been replaced by Ceph itself.)
- Update to version 12.1.4+git.1502871088.edf82a59df:
+ latest upstream luminous (a5f84b37668fc8e03165aaf5cbb380c78e4deba4)
* increment upstream version number to 12.1.4
- Update to version 12.1.3+git.1502698163.45fc586a51:
+ latest upstream luminous (6808af486d3bdce966232810004648502c967b13)
* increment upstream version number to 12.1.3
- Update to version 12.1.2+git.1501914637.59afe41c49:
+ spec file
* ceph-volume added to ceph-base RPM, where it co-exists with ceph-disk
- Update to version 12.1.2+git.1501865907.88dc8cdc9d:
+ latest upstream master (7350fa65d79b5f0ace427e609565f07be35cfe62)
* os/bluestore: allow multiple DeferredBatches in flight at once
(pr#16769, issue#20295, boo#1043399)
- Update to version 12.1.2+git.1501790037.82d6a9c37a:
+ osd: move CMPEXT response munging into completion function
(boo#1047244)
- Update to version 12.1.2+git.1501684697.aa6b0153ef:
+ upstream luminous v12.1.2 release candidate
+ latest upstream master (e27a78c9ed3984b54a312ced3ad67fcc8f464fc0)
* mon: add mgr metdata commands, and overall 'versions' command for all
daemon versions (boo#1050063, pr#16460)
- Update to version 12.1.1+git.1501669490.04d0738487:
+ mon: extensible output format for health checks
- Update to version 12.1.1+git.1501501507.6dd9e8fd8c:
+ update to latest upstream master (fa70335db25f2b661b329d5d8f37039b77f2d1ae)
+ spec file:
* build/ops: rpm: Drop legacy libxio support (pr#16449)
+ Other notable changes:
* osd: moved OpFinisher logic from OSDOp to OpContext (pr#16617)
- Update to version 12.1.1+git.1501231943.67577225bc
+ Notable changes:
* rgw_file: properly & |'d flags (boo#1047977)
- Update to version 12.1.1+git.1500447879.e5aac115d7:
+ upstream 12.1.1 (release candidate)
+ Other notable changes:
* build/ops: rpm: fix typo WTIH_BABELTRACE (pr#16366)
* build/ops: rpm: put mgr python build dependencies in make_check bcond
(pr#15940)
* ceph-disk: support osd new (pr#15432)
* ceph: allow '-' with -i and -o for stdin/stdout (pr#16359)
* cls/rbd: trash_list should be iterable (pr#16372)
* doc: update quick start to add a note for mgr create command for luminous+
builds (pr#16350)
* libradosstriper: fix format injection vulnerability (pr#15674)
* mgr/dashboard: add OSD list view (pr#16373)
* mon, osd: misc fixes (pr#16283)
* mon/MDSMonitor: fix segv when multiple MDSs raise same alert (pr#16302)
* mon/MgrMonitor: fix standby addition to mgrmap (pr#16397)
* mon/MgrMonitor: only induce mgr epoch shortly after mkfs (pr#16356)
* mon: Fix output text and doc (pr#16367)
* mon: add force-create-pg back (pr#16353)
* mon: skip crush smoke test when running under valgrind (pr#16346)
* os/bluestore: use reference to avoid string copy (pr#16364)
* osd: combine conditional statements (pr#16391)
* osd: restart boot process if waiting for luminous mons (pr#16341)
* rgw: S3 lifecycle now supports expiration date (pr#15807)
* rgw: acl grants num limit (pr#16291)
* rgw: delete object in error path (pr#16324)
* rgw: drop unused find_replacement() and some function docs (pr#16386)
* rgw: fix err when copy object in bucket with specified placement rule
(pr#15837)
* rgw: fix potential null pointer dereference in rgw_admin (pr#15667)
* rgw: raise debug level of RGWPostObj_ObjStore_S3::get_policy (pr#16203)
* rgw: req xml params size limitation error msg (pr#16310)
* rgw_file: add service map registration (pr#16251)
* test,rgw: fix rgw placement rule pool config option (pr#16380)
* test/fio: print all perfcounters rather than objectstore itself (pr#16339)
- Update to version 12.1.0+git.1500295948.39d2e5f424:
+ spec file:
* socat build dependency only for make check
* fix typo WTIH_BABELTRACE
- Update to version 12.1.0+git.1500284183.1ed97736ab:
+ Spec file changes:
* bump epoch (not relevant for SUSE)
* add socat buildrequires
* libcephfs2 now obsoletes libcephfs1 (boo#1047020)
* HAVE_BABELTRACE option renamed to WITH_BABELTRACE
* fix a regression in ceph-base %postun scriptlet
+ Selected other changes:
* Revert "msg/async: increase worker reference with local listen table enabled backend"
* build: build erasure-code isa lib without versions
* ceph-disk/ceph_disk/main.py: Replace ST_ISBLK() test by is_diskdevice()
* ceph.in: filter out audit from ceph -w
* common,config: OPT_FLOAT and OPT_DOUBLE output format in config show
* common/Timer: do not add event if already shutdown
* compressor/zlib: remove g_ceph_context/g_conf from compressor plugin
* core: make the conversion from wire error to host OS work
* core:" Stringify needs access to << before reference" src/include/stringify.h
* crush/CrushWrapper: make get_immediate_parent[_id] ignore per-class shadow hierarchy
* doc/mgr/dashboard: update dashboard docs to reflect new defaults
* doc/rados.8: add offset option for put command
* doc: fix link for ceph-mgr cephx authorization
* doc: update intro, quick start docs
* isa-l: update isa-l to v2.18
* librbd: fail IO request when exclusive lock cannot be obtained
* mgr,librados: service map
* mgr,mon: debug init and mgrdigest subscriptions
* mgr,mon: enable/disable mgr modules via 'ceph mgr module ...' commands
* mgr/ClusterState: do not mangle PGMap outside of Incremental
* mgr/restful: bind to :: and update docs
* mgr: Zabbix monitoring module
* mon,crush: create crush rules using device classes for replicated and ec pools via cli
* mon/MgrStatMonitor: do not crash on luminous dev version upgrades
* mon/MonClient: respect the priority in SRV RR
* mon/Monitor: recreate mon session if features changed
* mon/OSDMonitor: do not allow crush device classes until luminous
* mon/OSDMonitor: guard 'osd crush set-device-class'
* mon/mgr: add detail error infomation
* mon: Fix status output warning for mon_warn_osd_usage_min_max_delta
* mon: add support public_bind_addr option
* mon: collect mon metdata as part of the election
* mon: debug session feature tracking
* mon: do crushtool test with fork and timeout, but w/o exec of crushtool
* mon: fix hang on deprecated/removed 'pg set_\*full_ratio' commands
* mon: include device class in tree view; hide shadow hierarchy
* mon: maintain the "cluster" PerfCounters when using ceph-mgr
* mon: revamp health check/warning system
* msg/async: increase worker reference with local listen table enabled backend
* msg/async: make recv_stamp more precise
* msg: QueueStrategy::wait() joins all threads
* msg: make listen backlog an option, increase from 128 to 512
* os/blueStore: Failure retry for opening file
* os/bluestore/BlueFS: clean up log_writer aios from compaction
* os/bluestore/BlueFS: clear current log entrys before dump all fnode
* os/bluestore/KernelDevice: batch aio submit
* os/bluestore: Make BitmapFreelistManager kv itereator short lived
* os/bluestore: cap rocksdb cache size
* os/bluestore: clear up redundant size assignment in KernelDevice
* os/bluestore: differ default cache size for hdd/ssd backends
* os/bluestore: fix deferred_aio deadlock
* os/bluestore: move object exist in assign nid
* os/bluestore: only submit deferred if there is any
* os/bluestore: reduce some overhead for _do_clone_range() and _do_remove()
* os/bluestore: use bufferlist functions whenever possible
* os/filestore/FileJournal: FileJournal::open() close journal file before return error
* os/filestore/HashIndex: randomize split threshold by a configurable amount
* os/filestore: call committed_thru when no journal entries are replayed
* os/filestore: do not free event if not added
* os/filestore: queue ondisk completion before apply work
* os/filestore: require experimental flag for btrfs
* osd/OSD: auto class on osd start up
* osd/PrimaryLogPG solve cache tier osd high memory consumption
* osd/ReplicatedBackend: reset thread heartbeat after every omap entry
* osd/filestore: Revert "os/filestore: move ondisk in front
* osd: Check for and automatically repair object info soid during scrub
* osd: Preserve OSDOp information for historic ops
* osd: check queue_transaction return value
* osd: clear_queued_recovery() in on_shutdown()
* osd: compact osd feature
* pybind/mgr/dashboard: bind to :: by default
* pybind/mgr/dashboard: fix get kernel_version error
* rbd-mirror: ignore permission errors on rbd_mirroring object
* rbd: do not attempt to load key if auth is disabled
* rgw: fix compilation with cryptopp
* rgw,test: fix rgw placement rule pool config option
* rgw/rgw_frontend.h: Return negative value for empty uid in RGWLoadGenFrontend::init()
* rgw/rgw_rados: Remove duplicate calls in RGWRados::finalize()
* rgw: /info claims we do support Swift's accounts ACLs
* rgw: Correcting the condition in ceph_assert while parsing an AWS Principal
* rgw: Do not fetch bucket stats by default upon bucket listing
* rgw: Fix a bug that multipart upload may exceed the quota
* rgw: Initialize pointer fields
* rgw: fix leaks with incomplete multiparts
* rgw: fix memory leak in copy_obj_to_remote_dest
* rgw: fix not initialized vars which cause rgw crash with ec data pool
* rgw: fix radosgw-admin data sync run crash
* rgw: fix s3 object uploads with chunked transfers and v4 signatures
* rgw: fix the subdir without slash of s3 website url
* rgw: fix zone did't update realm_id when added to zonegroup
* rgw: log_meta only for more than one zone
* rgw: rgw_file skip policy read for virtual components
* rgw: using RGW_OBJ_NS_MULTIPART in check_bad_index_multipart
* rgw: verify md5 in post obj
* rgw_file: add compression interop to RGW NFS
* rgw：fix s3 aws v2 signature priority between header['X-Amz-Date'] and header['Date']
* vstart.sh: bind restful, dashboard to ::, not 127.0.0.1
* vstart: add --rgw_compression to set rgw compression plugin
* vstart: allow to start multiple radosgw when RGW=x
- Update to version 12.1.0+git.1499090920.2c11b88d14:
+ With this update, Ceph no longer autocreates a pool called "rbd" with id 0
Deployment tools must take this into account!
+ systemd: Add explicit Before=ceph.target (boo#1042973)
</description>
<summary>Recommended update for ceph</summary>
</patchinfo>
