I was wondering if anyone ever had a problem where they submitted a Proc SQL query in the morning that generated no error messages, to a DB2 server and it took minutes and reran the exact same query later and it ran for hours with an error message.
We use the libname approach.
We suspect that the SAS query changed but we would like to have more information as to how and why.
I've had identical queries run very quickly in the early morning and die totally in the afternoon. It's often a function of the load on the server and on the communications part of the network. I was using passthrough queries, so there wasn't any 'optimization' being done at the SAS end.
Another possibility is that a table was locked from reading and the query finally timed out.
Not much info to go on here - need to understand where SAS is executing (local, remote server, mainframe batch, etc.), and how about where the DB2 instance resides (OS platform, dedicated or shared server).
Any chance you might have the output for the SAS execution, which would provide some elapsed and CPU usage statistics, possibly? Something for your team to review in addition to asking questions about DB2 server (system or application) degradation at the time of the SAS execution.
At a minimum, sure, the DB2 server may be the bottleneck, serving your request, as well as others, or because of a overall system degradation condition. And, of course, it may be that the SAS program changed which could be contributing to the behavior/outcome.