Red Hat Training

A Red Hat training course is available for Red Hat Enterprise Linux

7.25. cluster

Updated cluster packages that fix several bugs and add two enhancements are now available for Red Hat Enterprise Linux 6.
The Red Hat Cluster Manager is a collection of technologies working together to provide data integrity and the ability to maintain application availability in the event of a failure.

Bug Fixes

Previously, the gfs2_convert utility or a certain corruption could introduce bogus values for the ondisk inode "di_goal_meta" field. Consequently, these bogus values could affect GFS2 block allocation, cause an EBADSLT error on such inodes, and could disallow the creation of new files in directories or new blocks in regular files. With this update, gfs2_convert calculates the correct values. The fsck.gfs2 utility now also has the capability to identify and fix incorrect inode goal values, and the described problems no longer occur.
The gfs2_quota, gfs2_tool, gfs2_grow, and gfs2_jadd utilities did not mount the gfs2 meta file system with the "context" mount option matching the "context" option used for mounting the parent gfs2 file system. Consequently, the affected gfs2 utilities failed with an error message "Device or resource busy" when run with SELinux enabled. The mentioned gfs2 utilities have been updated to pass the "context" mount option of the gfs2 file system to the meta file system, and they no longer fail when SELinux is enabled.
A race condition in the dlm_controld daemon could be triggered when reloading the configuration, which caused a dangling file pointer to be written to. Consequently, under certain rare conditions, dlm_controld could terminate unexpectedly with a segmentation fault, leaving Distributed Lock Manager (DLM) lockspaces unmanaged and requiring a system reboot to clear. This bug has been fixed, and dlm_controld no longer crashes when the configuration is updated.
Previously, errors generated while updating the resource-agents scheme were sometimes not reported. As a consequence, if an error occurred when updating the resource-agents schema, the update failed silently and later attempts to start the cman service could fail as well. With this update, schema errors are reported, and remedial action can be taken at upgrade time in case of problems.


The qdiskd daemon now automatically enables the master_wins mode when votes for the quorum disk default to 1 or when the number of votes is explicitly set to 1. As a result, quorum disk configuration is more consistent with the documentation, and a misconfiguration is avoided.
A new error message has been added to the qdiskd daemon, which prevents qdiskd from starting if it is configured with no heuristics in a cluster with three or more nodes. Heuristics are required in clusters with three or more nodes using a quorum device for correct operation in the event of a tie-break. Now, if no heuristics are specified and the cluster contains three or more nodes, the cman service fails to start and an error message is returned. This behavior prevents misconfigurations.
Users of cluster are advised to upgrade to these updated packages, which fix these bugs and add these enhancements.