RHBA-2011:0764 - Bug Fix Advisory

Topic

Updated corosync packages that add features and fix bugs are now available for
Red Hat Enterprise Linux 6.

Description

The corosync packages provide the Corosync Cluster Engine and C
Application Programming Interfaces (APIs) for Red Hat Enterprise Linux
cluster software.

This update fixes the following bugs:

• Multicast emulation caused an extra delay to the multicast packet

transmission, causing unnecessary retransmission of the packet. This
update adds the "miss_count_const" constant allowing the user to
specify how many times a message is checked before retransmission
occurs. (BZ#619496)

• When denied permissions from SELinux, corosync failed with a

segmentation fault. Corosync now passes an error
back to the API user when it is unable to create a connection between
the server and client instead of causing a segmentation fault.
(BZ#619918)

• When provided an invalid multicast address, corosync failed without

errors. This is now fixed, thus corosync displays an error when given
an invalid multicast address. (BZ#613836)

• Corosync client libraries delayed for 2 seconds before they displayed

an error on a shut down. This is now fixed, thus the exited flag value
before and after sem_wait is checked. If the value is true, ERR_LIBRARY
displays. (BZ#639023)

• The default TTL value in multicast was 1, preventing use on a routed

network. The TTL value is now configurable in the corosync
configuration file, thus multicast can now be used on a routed network.
(BZ#640311)

• BZ#640311 introduced a regression. (BZ#684920) (BZ#684930)
• Shared memory no longer is leaked if the corosync server unexpectedly

exits while connected to corosync clients. (BZ#665165)

• Running multiple instances of corosync simultaneously would succeed,

causing local node errors. This is now fixed to prevent initialization
of multiple instances of corosync. (BZ#626962)

• If cman ran the corosync init script, it would cause the corosync

init script to be blocked. This is now fixed to allow corosync to
create a Pid file and to allow cman to run corosync. (BZ#614104)

• Corosync was unable to capture system events and notify the user

about them. With this fix, SNMP MIB and daemon are added for system
event notification via DBUS and SNMP. (BZ#629380)

• Member objects in corosync were not found due to validation failure.

This is fixed with an addition to the objdb file, thus validation for
SNMP/DBUS integration is now successful. (BZ#675859)

• The corosync build contained invalid version information, which

caused rpmdiff to warn the user about version information changes.
This was fixed, thus pkgconfig files are now correctly configured to
display version as 1.2.3. (BZ#675741)

• Corosync rebuilds succeeded only on fresh installations due to a

regression issue. This is now fixed, thus corosync now rebuilds on
existing installations as well. (BZ#680258)

• A ring id file smaller than 8 bytes caused corosync to abort. This

was fixed by recreating the ring id file, thus corosync now does not
abort due to the ring id file. (BZ#675099)

• Inconsistent cluster.conf files amongst nodes caused a memory leak.

This is now fixed, thus a configuration reload via cman_tool no longer
causes a memory leak. (BZ#677975)

• During the recovery phase, aisexec exited unexpectedly, resulting in

a lost network token. This is now fixed, thus aisexec no longer exits
due to a lost token. (BZ#675783)

• UDPU transport is added, which simulates multicast via UDP unicast.

This adds a third transport option to broadcast and multicast in a
cluster. (BZ#568164)

• Fix abort that happens in rare circumstances during shutdown. (BZ#688691)

All users of corosync are advised to upgrade to these updated packages,
which resolve this issue.

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 High Availability for x86_64 6 x86_64
• Red Hat Enterprise Linux High Availability for x86_64 6 i386
• Red Hat Enterprise Linux Resilient Storage for x86_64 6 x86_64
• Red Hat Enterprise Linux Resilient Storage for x86_64 6 i386
• Red Hat Enterprise Linux High Availability (for RHEL Server) from RHUI 6 x86_64
• Red Hat Enterprise Linux High Availability (for RHEL Server) from RHUI 6 i386
• Red Hat Enterprise Linux Resilient Storage (for RHEL Server) from RHUI 6 x86_64
• Red Hat Enterprise Linux Resilient Storage (for RHEL Server) from RHUI 6 i386

Fixes

• BZ - 568164 - RFE add udpu communication transport to avoid requirement for multicast
• BZ - 613836 - invalid multicast address in corosync.conf file results in bad behavior
• BZ - 614104 - Starting or stopping corosync blocks cman from starting or stopping - corosync part
• BZ - 619496 - make corosync more resilient to delayed multicast packets
• BZ - 619918 - corosync should fail gracefully if selinux context is not available for an application using corosync apis
• BZ - 626962 - [RFE] don't allow multiple corosync processes to start on same system.
• BZ - 629380 - [RFE] Provide SNMP traps for corosync events
• BZ - 639023 - remove delay on normal shutdown of corosync server and return error to library user
• BZ - 640311 - Multicast TTL is 1 - prevents use on a multicast routed network
• BZ - 665165 - shared mem leak if corosync dies unexpectantly
• BZ - 675099 - when existing ring file is zero bytes, corosync aborts
• BZ - 675741 - pkgconfig files have wrong version information
• BZ - 675783 - token loss during recovery can trigger abort
• BZ - 675859 - missing upstream patches for full snmp/foghorn integration
• BZ - 677975 - Memory leak when objdb reload config
• BZ - 680258 - TPS rebuild problems with corosync
• BZ - 684928 - multicast ttl changed from 255 to 1 in defaults for totem
• BZ - 684930 - ttl is set for udpu when it should use kernel defaults
• BZ - 688691 - assertion failed "range < 16384" during corosync shutdown

CVEs

(none)

References

(none)