Rebooting Citrix Servers with RES Automation Manager

Sometimes we receive support calls from customers who report that their nightly reboot job for the Citrix servers was executed during business hours when users were already logged on to the servers. In most of these cases the cause of this behavior was a database or network problem at the time that the job should have run. Administrators came in the office in the morning, noticed that there was a problem and fixed it. Then the job executed with a delay of several hours.

Although RES Software has built several extra checks in the RES Automation Manager code to prevent this from happening there are a few thing you can do to further minimize the risk.

Launch window
In the Global Settings under Infrastructure > Datastore > Settings you can configure the Launch window. This is the time out period that RES Automation Manager uses to wait for Agents to pick up a job before it is canceled. The default value is 15 minutes and it should never exceed the amount of time between de reboot job and the start of the business day.

Conditions
On the reboot task itself you can configure conditions based on date & time. With these conditions you can specify that the task should never be executed between certain hours.

Run Books
Regardless of the time of day when the reboot job runs it is often better to reboot your Citrix servers in batches to insure that there are always servers available. You can, of course, schedule the same job more than once to different Teams or Agents but then you don’t have any dependencies. If the Agents from the first batch fail to come back online the second batch will start anyway, potentially with the same result. A RES Automation Manager Run Book does have such a dependency. You can divide your Citrix server into batches where the second batch will not start unless the first batch completed successfully. You can either create multiple jobs in the same Run Book to separate Teams or you can check the ‘Create multiple jobs and divide agents equally amongst them’ option and let RES Automation Manager do it automatically. In this example if the Team called Team contains all your Citrix servers they will be divided into three batches. Any new Team members will be added to one of the three batches. The option ‘Stop Run Book on error’ will prevent the next batch from rebooting should a job result in an error.