Analytics Software

Any input here would be greatly appreciated. Here is my situation:

We have a Scheduler Agent successfully installed and running various nightly jobs without problems (as a Windows service). However, we're seeing instances where internal IT is occasionally cycling our server in the early morning hours after maintenance operations. When this happens, our scheduled jobs stop.

What I think I've discovered via the agent log is that when the machine comes back up, all the Windows services (both DB, listener, and ODI Scheduler Agent) all restart as expected. But because the DB takes a couple minutes longer to start up and pair up with the listener, the Scheduler Agent which comes up almost immediately, has a few minutes where it can't find the DB.

To reiterate, the scheduler agent log will note a message like "Machine is rebooting." and "Wrapper stopped" indicating the shutdown. Those are immediately followed by some messages with timestamps a few minutes later of something like "Wrapper started as service" indicating the Scheduler Agent restarted when the box came back up, and then followed by a few connection attempt errors of the usual "SID not found" variety.

My theory is that the Scheduler Agent is coming back up much quicker than the DB, immediately starts pinging it and gets no response, and after a few rapid attempts gives up and disables itself. Though I'll admit there are no final messages in the log indicating it gave up and shut itself down...


Does anyone have any thoughts on how to prevent this? For example, is there a way to simply adjust the number of retry attempts that the Scheduler should try to find the database before giving up and disabling? It looks like I only need to but myself a couple minutes.

Thanks in advance-
JW

Edited by: jweicher on Jan 17, 2012 8:55 AM