This is crucial because it is the control script that actually starts up and shuts down our application. This step is commonly missed when changes to the control script are made. I have fielded many calls in the Response Center where an application is behaving differently on one node than another. Commonly, it is the fact that the control script was not distributed to one particular node. This seems like a "lame" excuse , but if you think about it, one node may be down due to urgent maintenance, and once rebooted, it is not uncommon to forget that you made a change to a control script for one of the many applications under the control of Serviceguard. This harks back to one of the cornerstones of high availability: well established and controlled IT processes. So remember, you have been warned .
Just before we proceed, you did remember to make all your scripts/programs, including your control file and application monitoring scripts, executable, didn't you?
