RHBA-2016:0313 - Bug Fix Advisory

Topic

Red Hat Ceph Storage 1.3.2 that fixes multiple bugs and adds various enhancements is now available.

Description

Red Hat Ceph Storage is a massively scalable, open, software-defined storage platform that combines the most stable version of the Ceph storage system with a Ceph management platform, deployment utilities, and support services.

Bug Fixes:

• The Ceph init script no longer calls the "ceph-disk" command on monitor nodes where the command is not present. As a result, an error is no longer displayed. (BZ#1225183)
• The librbd library did not resize format 1 image headers correctly in certain scenarios. Consequently, after resizing a RADOS block device (RBD) image, librbd operations returned the "librbd: image snap context is invalid!" error. With this update, librbd resizes the image headers correctly. (BZ#1275398)
• With this update, librbd has been updated to validate pools for RBD compatibility and automatically place new pools into self-managed snapshot mode. Now, when librbd is not able to force self-managed snapshot mode on a pool, an error is returned. (BZ#1276092)
• The Ceph monitor did not remove obsolete epochs after a compaction operation, causing the monitor store to grow to a large size on a disk, which could impact performance. Now, the monitor compacts the full epochs, thus improving performance. (BZ#1287947)
• Due to "last_epoch_started" mismatches, the OSD daemon could terminate unexpectedly with an assert when a placement group (PG) got stuck incomplete. Now, OSD skips the assert when the user sets the "osd_find_best_info_ignore_history_les" Boolean to "true". (BZ#1289313)
• With this update, librbd now properly reads diff files greater than 2 GB. (BZ#1289390)
• The rados(8) manual page now includes a description of the listwatchers command. (BZ#1293555)

Enhancements:

• With this update, a new "ceph osd crush tree" command has been added to Red Hat Ceph Storage. The command prints CRUSH buckets and items in a tree view. To learn more, see the Release Notes for Red Hat Ceph Storage 1.3.2. (BZ#1196613)
• The Ceph Object Gateway has been updated to support the "encoding-type" parameter when listing keys in a bucket. This change improves portability between the Amazon S3 public cloud object storage and the Ceph S3-compatible private cloud storage. (BZ#1220880)
• Due to a bug in previous version of the Ceph Object Gateway, S3 deletion operations left orphaned __shadow_ objects that used extra space. With this update, the "radosgw-admin" command has a new "orphans" subcommand, which finds and fixes the orphaned __shadow_ objects. (BZ#1254398)
• The Ceph Object Gateway has been updated to add a default quota configuration that is applied when no specific quota is enabled. As a result, users can now configure per-buckets and per-user defaults. The defaults are enabled if either max count or max size is enabled. (BZ#1259485)
• The Ceph monitor process now verifies that the monitor store is writable before participating in an election to avoid the possibility of losing cluster information during disk errors. (BZ#1262982)
• This update adds the ceph-monstore-tool utility that artificially inflates the monitor store leveldb database by using the "inflate-pgmap" command. This utility is useful for testing purposes, for example when the user needs to replicate scenarios in which the store significantly grows. (BZ#1263608)
• Support for modifying the size of the TCMalloc thread cache has been added. Increasing the thread cache size significantly improves Ceph cluster performance. To learn more, see the Release Notes for Red Hat Ceph Storage 1.3.2. (BZ#1297502)

Users of Red Hat Ceph Storage are advised to upgrade to these updated packages, which fix these bugs and add these enhancements.

Solution

Before applying this update, make sure all previously released errata
relevant to your system have been applied.

For details on how to apply this update, refer to:

https://access.redhat.com/articles/11258

Affected Products

• Red Hat Enterprise Linux Server 7 x86_64
• Red Hat Enterprise Linux Workstation 7 x86_64
• Red Hat Enterprise Linux Desktop 7 x86_64
• Red Hat Enterprise Linux for Scientific Computing 7 x86_64
• Red Hat Ceph Storage 1.3 x86_64
• Red Hat Ceph Storage Calamari 1.3 x86_64
• Red Hat Ceph Storage MON 1.3 x86_64
• Red Hat Ceph Storage OSD 1.3 x86_64

Fixes

• BZ - 1159756 - [RFE] SELinux Support for Ceph MON and OSD processes
• BZ - 1190201 - [RFE] SELinux support for calamari
• BZ - 1193228 - [RFE] SELinux support for RGW
• BZ - 1197045 - IOError: [Errno 30] Read-only file system: '/mnt/cephdeploy.conf'
• BZ - 1205927 - sudo ice_setup update when missing a parameter should report an error, and should not quietly do nothing.
• BZ - 1223731 - Parallel resize and write operation is causing the VM to crash on a RBD image, having object map feature enabled.
• BZ - 1224921 - IO operation is getting hanged with Mandatory exclusive lock enabled rbd image
• BZ - 1225183 - ceph sysvinit script calls ceph-disk when it may not be present
• BZ - 1228845 - rgw_frontends is unforgiving of whitespace and quotes.
• BZ - 1244158 - /etc/init.d/rbdmap relies on functions which are not defined in rhel 7
• BZ - 1245333 - rare leaked connection->session->message->connection cycle on client restart with blocked requests
• BZ - 1246223 - EC pool uses only min_size, not the isRecoverablePredicate
• BZ - 1250042 - RBD crash when we Force_off the VM after writeback blocked by IO errors
• BZ - 1251328 - backport #12523 (preserve rctx.handle in other RecoveryCtx constructor)
• BZ - 1252851 - [CACHE-TIER]: object not evicted from the cache pool based on target_max_object policy
• BZ - 1254398 - backport tool to identify orphan objects left behind by rgw deleting S3 objects
• BZ - 1255882 - Changing the size (number of replicas) in a pool should cause the pgs to have a new interval.
• BZ - 1261441 - ceph-deploy forgetkeys doesn't remove rgw keyring file
• BZ - 1262982 - RFE: mon verify that the store is writeable before participating in election
• BZ - 1268971 - Drop dependency on redhat-lsb-core
• BZ - 1282484 - [Ceph-deploy]: Ceph-deploy crashed while trying to remove a mon which is added manually
• BZ - 1288166 - [RHEL] package version dependencies are looser than needed
• BZ - 1289313 - osd/PG.cc: 288: FAILED assert(info.last_epoch_started >= info.history.last_epoch_started)
• BZ - 1293555 - [GSS] 'rados' man page doesn't mention the 'listwatchers' switch
• BZ - 1297502 - [RFE] Add support for modifying the TCMalloc thread cache
• BZ - 1298536 - ceph init script requires /lib/lsb/init-functions
• BZ - 1300829 - UBUNTU installation via ISO fails in step 'sudo ice_setup -d /mnt '
• BZ - 1301422 - UBUNTU installation - installation of ceph-test is failing
• BZ - 1302139 - VM permanently tries to read /dev/shm/lttng-ust-wait-5 saturating logs.
• BZ - 1302466 - HTTP 500 error, calamari cannot write to cthulhu.log
• BZ - 1302675 - OSD getting marked down while installing ceph-selinux package
• BZ - 1303241 - cthulhu crashes when osd_metadata is missing from the OSD_map
• BZ - 1303799 - [RHCeph 1.3.2 / 0.94.5-6] /var/lib/ceph Child dir secontext is unlabeled_t
• BZ - 1304032 - "orphans finish" command crashes: segmentation fault
• BZ - 1304341 - 1.3.2: Cannot open Calamari UI with selinux enforcing mode
• BZ - 1304454 - Seeing Error message in Vm log file 'libust[121880/121886]: Error: Error opening shm /lttng-ust-wait-5 (in get_wait_shm() at lttng-ust-comm.c:886)'
• BZ - 1304455 - Do not use lttng in rados/rbd by default
• BZ - 1305981 - [RHCeph 1.3.2] rgw logs in '/var/log/radosgw/

CVEs

(none)

References

• http://www.redhat.com/security/updates/classification/#normal