Tinkun wrote: > I have an odd problem starting jRLA on jbase 4.1.5.17 on AIX 5.3. > > We use the jDEMON script to start jbase with options StartOptionsRLA="- > ibs64000,2440,94" to increase locks. > Have you thought about trying to work out why your application requires so many locks? Or is this a third party application -looks like T24 to me. Please take the time to read through the posting guidelines as they would have told you to include a few more details about your application. 94 locks per group seems a strange number. It is not likely to fall on memory boundaries (but perhaps it does, I have not calculated it ). However, you are usually better with more groups to hash in to, to avoid any particular group exceeding the group count. I know that many T24 sites seem to have to increase lock allocations because of inefficient applications. You should monitor your live system for a while and see how high the watermark is for each group. > This works fine in our live environment but does not take effect on > our test server, which seems to only apply the default (Record locks: > 3020 locks maximum in 151 groups of 20 locks/group). > > Even if I start jRLA manually with jRLA ibs64000,2440,94 the setting > do not seem to take effect. > > Has anybody seen this before ? I noticed some other discussions but > no conclusive fix. > The following are likely:
1) The jRLA daemon was already running, but trying to start it again should have told you that; 2) When you start jRLA, it works out how much shared memory it is going to need for the lock tables. Your global and process limits for shared memory will apply here. It might be that you have not defined enough shared memory to accommodate 64000 locks. I no longer remember what happens in this case but it could be that it then tries to fall back on the defaults; 3) jRLA is picking up the settings from somewhere else and these are overriding your requests; Is that all the info? There are no messages when you start up jRLA manually and so on? Check the shared memory limits are the same on your test machine as they are on your live machine and see what else is using shared memory using the icps command. Jim --~--~---------~--~----~------------~-------~--~----~ Please read the posting guidelines at: http://groups.google.com/group/jBASE/web/Posting%20Guidelines IMPORTANT: Type T24: at the start of the subject line for questions specific to Globus/T24 To post, send email to [email protected] To unsubscribe, send email to [email protected] For more options, visit this group at http://groups.google.com/group/jBASE?hl=en -~----------~----~----~----~------~----~------~--~---
