4.149. rgmanager

Updated rgmanager packages that fix one bug are now available for Red Hat Enterprise Linux 5.
The rgmanager packages contain the Red Hat Resource Group Manager which provides the ability to create and manage high-availability server applications in the event of system downtime.

Bug Fix

BZ#827390
The filesystem and clusterfs resource agents were, in some cases, unable to unmount file systems that were exported using the nfsd utility. This update adds a new configuration option, "nfsrestart", as a workaround which ensures that the system is successfully unmounted. This new option is not compatible with the nfsserver resource agent, and requires the "force_unmount" option to be enabled.
All users of rgmanager are advised to upgrade to these updated packages, which fix this bug.
Updated rgmanager packages that fix one bug are now available for Red Hat Enterprise Linux 5.
The rgmanager packages contain the Red Hat Resource Group Manager which provides the ability to create and manage high-availability server applications in the event of system downtime.

Bug Fix

BZ#839195
Under rare conditions, rgmanager could attempt to free memory that had previously been freed. When this occurred, rgmanager terminated unexpectedly with a segmentation fault. This bug has been fixed and rgmanager no longer attempts to free previously-freed memory.
All users of rgmanager are advised to upgrade to these updated packages, which fix this bug.
Updated rgmanager packages that fix a bug are now available for Red Hat Enterprise Linux 5.
The rgmanager packages contain the Red Hat Resource Group Manager, which allows to create and manage high-availability server applications in the event of system downtime.

Bug Fix

BZ#852865
If the contents of the /proc/mounts file changed during a status check operation of the file system resource agent, the status check could incorrectly detect that a mount was missing. Consequently, a healthy service could have incorrectly be marked as having failed. This bug has been fixed and rgmanager's file system resource agent no longer reports false failures in the described scenario.
All users of rgmanager are advised to upgrade to these updated packages, which fix this bug.
Updated rgmanager packages that fix a bug are now available for Red Hat Enterprise Linux 5.
The rgmanager packages contain the Red Hat Resource Group Manager, which allows to create and manage high-availability server applications in the event of system downtime.

Bug Fix

BZ#859999
Due to a regression in the filesystem resource agent as part of the bug fix related to reading the /proc/mounts file, the filesystem resource agent status operations failed if the name of a file system resource contained the "/" character. This bug has been fixed and resources with "/" in their name no longer cause rgmanager to report failure for status operations.
All users of rgmanager are advised to upgrade to these updated packages, which fix this bug.
Updated rgmanager packages that fix a bug are now available for Red Hat Enterprise Linux 5.
The rgmanager packages contain the Red Hat Resource Group Manager, which allows to create and manage high-availability server applications in the event of system downtime.

Bug Fix

BZ#876962
When rgmanager received a remote start message for a particular service while already in the process of starting that service locally, a deadlock could occur. This sometimes happened during recovery of a service that had failed its start operation. This bug has been fixed and rgmanager works as expected.
All users of rgmanager are advised to upgrade to these updated packages, which fix this bug.
Updated rgmanager packages that fix several bugs are now available for Red Hat Enterprise Linux 5.
The rgmanager packages contain the Red Hat Resource Group Manager, which provides the ability to create and manage high-availability server applications in the event of system downtime.

Bug Fixes

BZ#693855
A mirror device failure during the relocation of the High Availability LVM service (HA-LVM) could cause, under certain circumstances, the service to fail. This bug has been fixed and now the mirror device failure no longer affects the HA-LVM service in such a case.
BZ#723819
The orainstance.sh resource agent did not detect all startup failures properly. The underlying source code has been modified and all failures are now detected correctly.
BZ#756180
LVM resource agent could not update logical volume tags if there were missing physical volumes. This bug has been fixed and the logical volume tags are forcibly removed if the physical volumes are missing.
BZ#769730
If the cman service was stopped while the rgmanager service was running, rgmanager sometimes exited uncleanly without releasing its Distributed Lock Manager (DLM) lock space. Consequently, it was impossible to shut down rgmanager and cman. Now if the user mistakenly attempts to stop cman service while rgmanager is still running, rgmanager no longer stops in this situation.
BZ#773372
If the /etc/lvm/lvm.conf file was changed after the last initrd (initial ramdisk) rebuild, the LVM resource agent failed. This agent has been modified to generate a warning message and no longer fails in such a case.
BZ#789366
If a service with a relocate failover policy failed and the relocation operation failed as well, the service could be restarted locally. Due to an error in the source code, the service afterwards stopped, even if the local restart succeeded. This error has been fixed, and these services no longer stop after a successful local restart.
BZ#819595
When the root file system was full, rgmanager randomly killed applications when trying to force-unmount. The underlying source code has been modified and applications are stopped instead of killed in this case.
BZ#820632
Under rare conditions, rgmanager attempted to free memory that had been previously freed. As a consequence, rgmanager terminated unexpectedly with a segmentation fault. This bug has been fixed and rgmanager no longer attempts to free previously-freed memory.
BZ#834459
When rgmanager received a remote start message for a particular service while already in the process of starting that service locally, a deadlock could occur. This sometimes happened during recovery of a service that had failed its start operation. This bug has been fixed and rgmanager works as expected.
BZ#847125
If the contents of the /proc/mounts file changed during a status check operation of the file system resource agent, the status check could incorrectly detect that a mount was missing and mark a service as failed. This bug has been fixed and rgmanager's file system resource agent no longer reports false failures in the described scenario.

Enhancements

BZ#819494
A new "prefer_interface" parameter has been added to the rgmanager ip.sh resource agent. This parameter is used for adding an IP address to a particular network interface if a cluster node has multiple active interfaces that have IP addresses on the same subnetwork.
BZ#822066
In some cases, "fs unmount" command and clustersfs resource agents were unable to unmount the file systems which were exported by the nfsd utility. The new nfsrestart option to enable a last resort workaround prior to failing to unmount the file system has been added. The new option requires force_unmount="" to be enabled and it is not compatible with nfsserver resource agent.
All users of rgmanager should upgrade to these updated packages, which fix these bugs and add these enhancements.