4.12. clustermon

Bug Fix

BZ#958026

Prior to this update, the dynamic library that represents the CIM provider of a cluster status was not built with all the required dependencies, and certain symbols could not be resolved. As a consequence, the cluster status could not be accessed by CIM. This update adds the missing dependencies to the dynamic library, and the cluster status is now accessible as expected.

Bug Fixes

BZ#847289

Previously, the SNMP agent exposing the cluster status and shipped as cluster-snmp caused the SNMP server (snmpd) to terminate unexpectedly with a segmentation fault when this module was loaded, and the containing server was told to reload. This was caused by an improper disposal of the resources facilitated by this server, alarms in particular. Now, the module will properly clean up such resources when being unloaded, preventing the crash on reload.

BZ#882277

Prior to this update, the dynamic library that represents the CIM provider of a cluster status was not built with all the required dependencies, and certain symbols could not be resolved. As a consequence, the cluster status could not be accessed by CIM. This update adds the missing dependencies to the dynamic library, and the cluster status is now accessible as expected.

BZ#957798

When modclusterd was about to associate the local machine with a particular cluster node entry from the cluster configuration, it first tried CMAN API in an improper way, yielding no expected results. This process occurred every five seconds. As a consequence, it had to periodically resort to iterative detection through local interface addresses. Also when logging for CMAN enabled, and including membership messages, it would cause messages arising from the CMAN API misuse to be emitted. Now, the API is used as expected, which corrects the aforementioned consequences.

BZ#965792

Previously, a segmentation fault occurred in the modclusterd daemon and/or modcluster, ricci cluster-dedicated module, when processing large cluster.conf files (or peer-delivered counterparts). This was caused by the allocation of large amounts of memory not available on the stack. With this update, a patch has been introduced to allocate memory on the heap, and provide an error if not enough memory is available. As a result, crashes no longer occur in the described scenario.