Joined: 23 Nov 2006 Posts: 19270 Location: Inside the Matrix
Does anyone know of a situation where a batch job using some plan would "lock" that plan (or a table referenced by that plan/dbrm) causing a -904 in the bind step when a compile/link/bind was run?
One of the people from "down the hall" came in this afternoon and said that a long-running batch job was preventing them from binding programs that used tables that were in use by the batch job. When the job was canceled, they were able to bind. The job did no updates, only reads against 2 tables.
They were unable to get any satisfactory answer from the db2 support people and asked if i might be able to help find a resolution. Their concern is that as both compiling and testing is increasing, this may become a major problem if it is not resolved.
I can probably get more specifics if needed, but for starters, i was just looking to see if this is a situation anyone has had to resolve. Searching our forum and then the web did not find many reports of this (there are a few, but very few). My guess is that there is come kind of configuration problem. . .
Joined: 14 Jul 2008 Posts: 1249 Location: Richfield, MN, USA
You're the answer man Dick -- what're you doing asking a question? Seriously, I was involved in a project one time where many developers were compiling, linking, and binding DB2 programs all day long. It wasn't unusual for us to run across contention problems because of this. I think it was the SYSPLAN table that got hit often. We'd just resubmit our compiles and be on our way. Sorry I couldn't be of more help.