Veritas Cluster Server KM for PATROL

Release Notes for v2.5.00

Home  Previous  Next

What's New

Fully compatible with BMC ProactiveNet Central Monitoring Administration, a feature supported by BMC ProactiveNet Performance Management version 9.0.00.

Veritas Cluster Server KM for PATROL now supports Veritas Storage Foundation v6.x.

In order to easily identify which step the initial discovery is processing, the setup icon now reflects the successive phases:

Upgrading: this information is only available for PATROL Classic Console.

Loading: indicates that commands are being loaded.

Initializing: indicates that KM paths are being initialized. Missing paths are now specified in the system Output Window.

Discovering: indicates that the application is being monitored.

Cleaning: indicates that the temp files are being cleaned.

Validating: indicates that the application paths are being cleaned.

Veritas Global Clusters support: The KM can now discover and monitor shared global customer service groups and resources. It can also fail over between global cluster nodes following the online shared service group. For that purpose, new application classes were added for monitoring global cluster heartbeats over WAN.

Veritas Cluster Server KM for PATROL now checks the main.cf file and triggers a warning if an error is detected in the VCS configuration. The default VCS configuration directory path can be changed using the Paths KM commands.

Parameters

Two new parameters were added to the VCS_RESOURCE application class, VCSResourceState and VCSResourceStatus, to monitor the enabled/disabled state of the resource.

The parameter VCSSystemLoadWarnDuration was added to the VCS_SLAVE application class to indicate how long the system load has been over the warning threshold set in Veritas Cluster Server with the LoadTimeCounter system attribute.

Three new parameters were added to the VCS_GSYSTEM application class:

VCSGSystemAutoDisabled: to indicate whether the service group has been auto-disabled on the system by Veritas Cluster Server.

VCSGSystemEnabled: to indicate whether the service group has been enabled on the system.

VCSGSystemProbed: to indicate whether the service group has been successfully probed on the system.

The VCSRSystemProbed parameter was added to the VCS_RSYSTEM application class to indicate whether the resource has been successfully probed on the system.

Changes and Improvements

The KM performance has been improved.

All collectors are now automatically refreshed as soon as the KM configuration and the initial discovery are complete.

The default number of instances of Application Classes has been removed. This information must be defined by the user. When set to 0, the monitoring of a specific Application Class is not performed.

Logs: The default log scan limit was increased to 500 KBytes.

When one or more required paths are missing, the setup instance label will now indicate which paths are missing. A message will also be displayed in the System Output Window.

The VCSRemoveTempFiles parameter was removed.

All previous pconfig variables available under /VCS are now split to /VCS and /Runtime/VCS while upgrading the KM.

Debug Mode:

The time during which the debug information is logged can now be set by the user.

The KM-related messages that are sent to the System Output Window (SOW) are also logged to the VCS.log file available under <PATROL_HOME>/log.

The KM status report was improved to include more performance details. It is also more regularly updated and indicates when the last collection/discovery was performed.

GAB Monitoring: All GAB ports are discovered as soon as they are registered/reported and monitored throughout the agent restart. The KM will therefore continue to monitor ports that had been previously detected. If a previously detected port does not appear anymore in the GAB configuration (gabconfig -a), the port state (VCSGabState) will be set to 3 = Not Registered, which will trigger an alarm through the VCSGabStatus parameter.

The KM failover time has been reduced to sooner discover VCS node status errors.

The polling interval can now be modified for the server discovery to sooner detect server changes, cluster fail over, and node type changes.  

Fixed Issues

With some Windows platforms the Debug KM Command was prompting for an OS login to turn the debug on.

The VCSDaemonState parameter was set to "Not Running" when the PATROL Agent could not fork processes, which resulted in invalid alarms on the VCSDaemonStatus parameter.

Veritas Cluster Server KM for PATROL crashed when attempting to trigger events with more than 15 fields when installed on Windows 2008 R2 platforms with PATROL Agent 3.8.50.

Veritas Storage Foundation v6.x:

The VCSSystemFrozenState parameter was not properly set when the system was in frozen state since this parameter failed to detect the TFrozen state.

The VCSGSystemState parameter was not properly set.