RHBA-2011:0036 - Bug Fix Advisory

Topic

Updated cman packages that fix bugs and add enhancements are now available.

Description

The Cluster Manager (cman) utility provides user-level services for
managing a Linux cluster.

This update applies fixes for the following bugs:

• gfs_controld: fix plock owner in unmount. (BZ#624554)
• gfs_controld: fix plock owner syncing. (BZ#624156)
• fenced: use post_join_delay after cluster join. (BZ#575952)
• group_tool: list more than 64 groups. (BZ#561509)
• Add extended information to 'group_tool dump plocks' (gfs). (BZ#514264)
• fence_ipmilan now supports HP iLO 3 devices. (BZ#548575)
• mkqdisk debug option is no longer position dependent. (BZ#555246)
• cman can now be rebuilt with latest NSS libraries. (BZ#575157)
• gfs_controld: set last_plock_time for ownership operations. (BZ#578632)
• fence_scsi get_scsi_devices() has vastly improved performance. (BZ#564468)
• If ccsd fails to bind to a socket or send to a socket, ccsd fails to report

the IP address and port. (BZ#573996)

• fence_ilo throws an exception if the user does not have correct privileges.

(BZ#576176)

• fence_ack_manual man page is now up to date. (BZ#578604)
• Fencing agents now return correct return codes. (BZ#583034)
• Standard output is now closed in fence_rsb. (BZ#583040)
• fence_ilo no longer depends on OpenSSL. (BZ#583049)
• Obtaining metadata from fencing agents now works properly. (BZ#583053)
• libcman no longer leaks file descriptors. (BZ#585218)
• fail_to_recv_const has been changed to 2500. (BZ#587080)
• Fence_vmware and fence_vmware_helper syntax errors have been fixed.

(BZ#590304)

• Quorum is no longer recalculated after a node abruptly leaves cluster.

(BZ#590304)

• fence_apc no longer fails for some port numbers. (BZ#606315)
• Consensus timeout is now optimized based on node count at cluster startup,

significantly reducing failure recovery time in clusters with two members.
(BZ#611391)

• Qdiskd now gives up on heuristics that do not respond prior to the qdisk

timeout. (BZ#589266)

• Multiple mount attempts when mounting the same gfs2 file system no longer fail

incorrectly. (BZ#559735)

• Fencing agent parameters now appear in agent metadata output. (BZ#619034)
• 'port' is now a synonym of 'module_name' for fence_drac5. (BZ#619776)
• Some fencing agents do not support login using keys. (BZ#582334)
• The package no longer overwrites users' log files in /var/log/cluster

(BZ#659090)

In addition, this update adds the following enhancements:

• Use of clustering is now supported within guest virtual machines managed by

Red Hat Enterprise Virtualization (BZ#595458)

• fence_scsi_test can now test preempt-and-abort, enabling administrators to

better decide whether their arrays support fence_scsi. (BZ#603838)

• The use of broacast mode is now supported in some configurations. (BZ#629652)
• Fence-Agents: Support non-default TCP ports in fence_wti for the different

services (SSH / Telnet). (BZ#572996)

• Two-node clusters may now delay fencing to resolve race-to-fence situations.

(BZ#613064)

• IBM IMMv2 devices are now supported by fence_ipmilan. (BZ#614198)
• Adding LUNs to fence_scsi without restarting the cluster is now supported.

(BZ#616138)

• A fencing agent for use with Cisco Unified Computing System environments is

provided. (BZ#579859)

• A fencing agent for use with Red Hat Enterprise Virtualization Manager is

provided. (BZ#595835)

• A fencing agent for use in clusters utilizing certain switches in conjunction

with iSCSI storage is provided. (BZ#572863)

• WTI VMR devices no longer require their prompts to be set to "IPS".

(BZ#578615)

All cman users are advised to upgrade to these updated packages, which fixes
these issues and add these enhancements.

Solution

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

This update is available via the Red Hat Network. Details on how to
use the Red Hat Network to apply this update are available at
http://kbase.redhat.com/faq/docs/DOC-11259

Affected Products

• Red Hat Enterprise Linux Server 5 x86_64
• Red Hat Enterprise Linux Server 5 ia64
• Red Hat Enterprise Linux Server 5 i386
• Red Hat Enterprise Linux Workstation 5 x86_64
• Red Hat Enterprise Linux Workstation 5 i386
• Red Hat Enterprise Linux for Power, big endian 5 ppc
• Red Hat Enterprise Linux Server from RHUI 5 x86_64
• Red Hat Enterprise Linux Server from RHUI 5 i386

Fixes

• BZ - 514264 - Add extended information to 'group_tool dump plocks' (gfs)
• BZ - 555246 - mkqdisk debug option (-d) specified position
• BZ - 559735 - GFS2 mount fails incorrectly after correctly failed second-mount attempt
• BZ - 561509 - group_tool only shows 64 groups
• BZ - 564468 - fence_scsi get_scsi_devices() too slow, maybe hanging
• BZ - 572996 - Fence-Agents: fence_wti, support non-default TCP ports for the different services (SSH, Telnet)
• BZ - 573996 - ccsd does not log what IPs are used for binding/sending packets to
• BZ - 575157 - cman cannot be rebuilt with latest nss libraries
• BZ - 575952 - fenced post_join_delay doesn't help after quorum is restored
• BZ - 576176 - fence_ilo will throw exception if user does not have power priviledges
• BZ - 578604 - fence_ack_manual man page is out of date
• BZ - 578615 - Add direct support for WTI VMR
• BZ - 578632 - gfs_controld: set last_plock_time for ownership operations
• BZ - 579859 - [Cisco 5.6 feat] Include Cisco UCS Fencing Agent
• BZ - 582334 - Some fencing agents do not support login using keys
• BZ - 583034 - Return codes of fencing agent
• BZ - 583040 - Force stdout close for fence_rsb
• BZ - 583049 - Remove dependency on OpenSSL in fence_ilo
• BZ - 589266 - qdisk should have a timer for heuristic "programs" instead of relying on "program" to provide timeout
• BZ - 590304 - cman-2.0.115-34.el5 /sbin/fence_vmware and /sbin/fence_vmware_helper
• BZ - 595385 - Provide fence-rhev agent that uses the RHEV REST API
• BZ - 595394 - Qdisk: quorum is recalculated after node abruptly leaves cluster
• BZ - 595458 - Full support/testing for RHCS on RHEV guests
• BZ - 606315 - fence_apc fails for some port numbers
• BZ - 613064 - Method to cause one node to delay fencing in a two node cluster
• BZ - 616138 - Ability to add new LUNs to registration for a node w/o needing to restart cluster
• BZ - 619034 - Not all parameters appear in metadata
• BZ - 619776 - make "port" a synonym of "module_name" for drac5
• BZ - 624156 - gfs_controld: fix plock owner syncing
• BZ - 624554 - gfs_controld: fix plock owner in unmount
• BZ - 629652 - Provide support for broadcast mode (vs. multicast) for a limited set of cluster configurations
• BZ - 642237 - Testing of fence_brocade switches against FOS versions 6.1.x, 6.2.x, 6.3.x, 6.4.x
• BZ - 659090 - ownership of /var/log/cluster clobbers log files

CVEs

(none)

References

(none)