Manual restart of a VCS managed Application type resource may not initiate a VCS failover
Article ID: 100021776
Updated On:
Description
Application cron_res (User = rootStartProgram = "/var/tmp/crontabs/cron-program start"StopProgram = "/var/tmp/crontabs/cron-program stop"CleanProgram = "/var/tmp/crontabs/cron-program clean"MonitorProgram = "/var/tmp/crontabs/cron-program monitor")
The following shows the monitor process intervals for the Application agent:
#hatype -display Application |grep MonitorIntervalApplication MonitorInterval 60
If the above cron_res Application is manually restarted by executing the StopProgram followed immediately by the StartProgram, and if VCS did not perform the monitor during that restart time period, no fault or status change is detected.
Resolution
VCS monitors managed application resource states by executing a "Monitor" process on a regular interval. The monitor process is designed to query the managed resource in such as way as to confirm that the resource is online and responding. The default for this behaviour for many resource types is to perform the monitor process once every 60 seconds (this value can be customized).
If a managed resource is manually offlined and restarted using non-VCS managed controls and the application completes startup before the next monitor process is performed VCS will not detect any status change for that resource.
The following is an example using an Application type resource used to manage a Cron program:
Issue/Introduction
