RHBA-2012:0898 - Bug Fix Advisory

Topic

Updated ricci packages that fix multiple bugs and add two enhancements are now
available for Red Hat Enterprise Linux 6.

Description

The ricci packages contain a daemon and a client for remote configuring and
managing of clusters.

This update fixes the following bugs:

• Previously, the ccs utility was not able to configure IPV6 hosts. The ccs

utility has been updated to support the IPV6 protocol, so that it can be used to
configure IPV6 hosts. (BZ#724014)

• The cman and ccs utilities contain different cluster schemas for Red Hat

Enterprise Linux 6.2. A different cluster schema on a node from the schema
contained in ccs could cause a valid cluster.conf file to be interpreted as
invalid by ccs. A new ccs option, "--getschema", has been added. Using this
option, if cluster schemas are different between what is included in ccs and the
information on a node, the cluster schema on the node is used. (BZ#726772)

• Previously, the ccs utility did not check to see if the cluster.conf file was

invalid. Therefore, when parsing an invalid cluster.conf file, ccs could
terminate unexpectedly followed by a traceback, however with no detailed
information about the problem provided. With this update, ccs checks to see
whether cluster.conf is valid; if the file is invalid, ccs prints a helpful
error message and exits gracefully. (BZ#731113)

• Previously, the "ccs_sync" command did not return a non-zero exit code if an

error occurred or the ricci daemon was not running, even when running the
command with the "-w" option (to exit with a failure status if any warnings were
issued). The underlying source code has been modified so that "ccs_sync" with
the "-w" option now returns "1" on failure. (BZ#738008)

• When running the "ccs" command with the "--checkconf" and "-f" options, ccs

verifies that all the nodes in the file specified contain the same cluster.conf
file. The XML code in the configuration file is indented whereas the XML code in
the live configuration file from ricci is not, and therefore bare comparison of
such strings always failed. As a consequence, ccs did not verify each node in
the local cluster.conf file. The comparison method has been improved so that
"ccs --checkconf -f" now correctly verifies whether all the nodes in the file
contain the same cluster.conf file. (BZ#738567)

• Previously, if the user installed the ricci packages but did not install the

modcluster package, any attempt to run cluster commands using ricci failed. With
this update, users are no longer allowed to install only the ricci packages; the
modcluster package is now required as a dependency. As a result, cluster
commands can be executed as expected. (BZ#742345)

• Previously, when the user configured a virtual machine service, the virtual

machine was not displayed in the output of the "ccs --lsservices" command. The
ccs utility has been modified to specifically check for virtual machine services
in the configuration file. Now, running "ccs" with the "--lsservices" option
prints the proper output. (BZ#770637)

• Previously, if the user installed ricci and did not set a user password, the

user was unable to connect to ricci. This could lead to confusion, because the
user was not prompted to provide the password and therefore not aware of the
requirement. Also, no explanation was logged in the system log. With this
update, ricci logs a warning message to syslog on startup if the ricci user
password is not set. (BZ#773383)

This update also adds the following enhancements:

• Prior to this update, the ricci daemon processed tasks, but did not log any

information about when the tasks were run. With this update, when ricci is asked
to spawn a worker process, the date and time information is now logged using
syslog. (BZ#738797)

• Prior to this update, the ccs utility did not provide a way to configure

Redundant Ring Protocol (RRP). This update adds additional configuration options
for RRP into the ccs utility so that users can configure RRP using ccs.
(BZ#758823)

All users of ricci 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.

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

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 - 721113 - ccs should not backtrace when clustermon is not running
• BZ - 724014 - Allow ccs to connect to ipv6 hosts
• BZ - 726772 - cman and ccs packages contain different cluster.rng schema files for RHEL 6.3 Cluster
• BZ - 729011 - Failed to find the .debug_info section in a debuginfo file
• BZ - 729168 - Improve ccs documentation for --start/--stop
• BZ - 731113 - ccs does not provide a useful error message when it parses a bad cluster.conf
• BZ - 734910 - ccs should not traceback when passing a directory with '-f'
• BZ - 738008 - ccs_sync exit code always zero
• BZ - 738567 - ccs --checkconf works improperly when used on cluster.conf file
• BZ - 758823 - cman and corosync RRP handling are not consistent and needs improvements
• BZ - 797267 - ricci doesn't properly return the cluster schema
• BZ - 797292 - ccs --createcluster should verify that a cluster.conf doesn't exist before creating one

CVEs

(none)

References

(none)