RHBA-2014:1526 - Bug Fix Advisory

Topic

Updated pcs packages that fix several bugs and add one enhancement are now
available for Red Hat Enterprise Linux 6.

Description

The pcs packages provide a command-line tool to configure and manage the
Pacemaker and Corosync tools.

This update fixes the following bugs:

• The pcs utility was using an incorrect location to search for cluster node

names, and the "pcs cluster standby" command therefore could not find the
specified cluster node. As a a consequence, it was not possible to put cluster
nodes in standby mode. With this update, pcs properly searches for node names in
the /etc/cluster/cluster.conf file and putting cluster nodes in standby mode
works correctly. (BZ#1029129)

• Previously, pcs was unable to create user ID (UID) and group ID (GID) entries

in the cluster.conf file. Consequently, non-root users could not gain access to
corosync.conf and therefore could not access corosync. Now, pcs support for
configuring UID and GID entries in cluster.conf has been added, and non-root
users can be granted access to corosync with the "pcs cluster uidgid" command.
(BZ#1025054)

• When using the "pcs resource create" command with the "--group" option, pcs

created a resource and added it to a resource group in two separate steps
instead of one. Consequently, a resource that was added to a pre-existing
resource group sometimes initially started on an incorrect node. With this
update, pcs creates a resource in a resource group as a single step when
"--group" is used, and the created resource starts on the correct node.
(BZ#1066927)

• When adding a STONITH fencing level, pcs sometimes mistakenly detected that

some nodes were not a part of a cluster . As a consequence, it was not possible
to add a STONITH level unless the "--force" option was used. Now, pcs correctly
determines whether a node is a part of a cluster and, as long as a valid node is
used, adding a pcs STONITH level no longer requires the use of the "--force"
option. (BZ#1019410)

• Previously, pcs did not allow the use of the following attributes for STONITH

fencing agents: pcmk_reboot_action, pcmk_monitor_action, and pcmk_status_action.
As a consequence, using any of these attributes when configuring a STONITH agent
caused the configuration to fail. With this update, pcs correctly allows the use
of the attributes and the configuration no longer fails when they are used.
(BZ#1094517)

• The pcs utility did not properly handle clones of a group when removing

resources from the cloned group. As a consequence, the "pcs resource unclone"
and "pcs resource delete" commands removed only the first resource in a group
when they were supposed to remove the entire resource group. With this update,
pcs handles resources in cloned groups correctly and removing cloned resource
groups works as expected. (BZ#1108778, BZ#1107612)

• Due to an error in detecting whether resource groups are managed, pcs

sometimes could not delete a cloned resource group or a master or slave resource
group. With this update, pcs detects the status of resource groups correctly,
and deleting the mentioned resource groups proceeds normally. (BZ#1107965)

• Previously, pcs attempted to use the corosync.conf file when listing cluster

nodes using the "pcs status nodes corosync" command. However, corosync.conf does
not exist in Red Hat Enterprise Linux 6. As a consequence, "pcs status nodes
corosync" failed to execute. Now, cluster nodes on Red Hat Enterprise Linux 6
are listed using the cman_tool program and the /etc/cluster/cluster.conf file,
and "pcs status nodes corosync" functions correctly. (BZ#1094812)

In addition, this update adds the following enhancement:

• The pcsd daemon has been added to the pacemaker packages for Red Hat

Enterprise Linux 6, which allows users to remotely start, stop, enable, or
disable the cluster, and also to remotely set up the cluster configuration.
(BZ#1035300)

Users of pcs are advised to upgrade to these updated packages, which fix these
bugs and add this enhancement.

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/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 - 896628 - [RFE] Implement configuration syncing over all nodes [rhel6]
• BZ - 896632 - [RFE] Give pcs and option to start/stop all cluster nodes [rhel6]
• BZ - 1019410 - 'pcs stonith level add' gives "Error: <node> is not currently a node (use --force to override)"
• BZ - 1025054 - Need ability to create uidgid entries for corosync in cluster.conf using pcs
• BZ - 1026431 - pcs incorrectly detects RHEL6 variants
• BZ - 1029129 - pcs cluster standby nodename doesn't work
• BZ - 1033538 - pcs does not correctly display status for cloned resource groups
• BZ - 1038107 - "pcs cluster enable/disable" do not sets correctly the startup of the daemons
• BZ - 1091063 - unable to delete clone resource with global-unique attr set
• BZ - 1091066 - pcs incorrect option parsing
• BZ - 1094812 - 'pcs status nodes corosync' gives an error instead of providing nodes info
• BZ - 1102444 - 'pcs stonith list' -> Error: no metadata for /usr/sbin/fence_check (and others)
• BZ - 1104652 - Unable to delete resources in offline mode due to pcs trying to stop a service
• BZ - 1107612 - 'resource delete' behaviour confusing for clone of group with more than one resource
• BZ - 1107965 - Cloned or M/S group cannot be deleted
• BZ - 1108778 - 'pcs resource unclone group-clone' removes all the group resources but the first one
• BZ - 1112727 - Need ability to upgrade cib to use ACL support in existing 6.5 pacemaker cluster

CVEs

(none)

References

(none)