Is there a way to force MQ on the distributed platform to take a checkpoint rather than waiting for 10,000 operations to occur, which is when checkpoints are normally taken? Or, can this value be tuned to a lower number than 10,000?
Thanks - Tim -----Original Message----- From: Paul Clarke [mailto:paulg_clarke@;UK.IBM.COM] Sent: Thursday, October 31, 2002 2:04 PM To: [EMAIL PROTECTED] Subject: Re: Relation of currdepth to disk space >The space will not get freed up until either a) you delete and redefine the >queue, or b) the QMGR is stopped and restarted. This is not quite true, we do indeed release disk space at these times but these aren't the only times... The others are :- 1) At a checkpoint If checkpoint sees the same queue empty multiple times and the queue is using a significant amount of disk space then the queue is truncated. 2) When the queue is unloaded When a queue isn't open by any application (including internal MQ 'applications' e.g. MCA) then it is scheduled for unload. The queue will be unloaded the second time a checkpoint sees the queue scheduled for unload. 3) When we take a media image of the queue We compact a local queue prior to taking a media image. I imagine the apparent arbitrary nature that was referred to in the earlier append is just the moment when the queue became unused or when a checkpoint was taken. Cheers, P. Paul G Clarke WebSphere MQ Development IBM Rochester,MN 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 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