So, we're a new install (769 Solaris), and I've been getting bunch of emails with the subject "agent has crashed (tscomid=N)". How can I figure out why these agents are crashing? All of the stderr.log files are empty.

This message has been edited. Last edited by: mamercad,

1. identify a request that consistently crashes the agent.
2. see to it that you are the only person using the server.
3. switch on tracing, run the request, and switch tracing off.
4. see if you can discern anything from the trace files - since the agent crashes there is little chance that you indeed find something, but it is worth the try.
5. try to find out at what point in the procedure the crash occurs - do this by progressively removing bits of code from the request until you reach a point where you find the statement or set of statements that is/are the culprit.

I would do steps 3 and 4 only if everything else fails, reading and interpreting the trace files can be very time-consuming.

Thank you, I will try these tips.