Hi all, A client of ours is having an issue with ridiculously long commit times, and I was wondering if anyone else has experienced this behaviour. It goes a little like this...
1) A batch job runs which inserts a number of (200 or so) records into a DB2 table, under 1 unit of work. 2) Each insert into the DB2 table triggers a WLM controlled COBOL stored procedure. The procedure has linked CSQBRSTB. 3) The stored procedure makes some MQI calls to put a message to a queue. 4) The batch job issues a commit after all the inserts have been actioned. 5) Then, the batch job waits for 5 minutes for the commits to complete. Watching the queue, the messages are being committed in groups of 20 or 30 messages spaced over the 5 minutes. We're on z/OS 1.2, MQ 2.1 and DB2 7.1, and we've also tested this with MQ5.3, similar results. Seem familiar to anyone? I'm wondering if this is even a valid way of doing RRS controlled coordination? Ian Metcalfe, EDS Australia Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
