After a whole production day running James 2.1 after having set <deliveryThreads> to 1, the CPU problem has never come back again! And no FileNotFoundException in the outgoing mail (the latter were very few anyway).
I still got ConcurrentModificationException 15 times, but yesterday I got it 190 times during the same hours. It sounds like the defect fixed by Noel today in the v2.1.1a5 test build was amplified by multiple threads concurrency. I will (1) run (tomorrow or Monday) v2.1.1a5 with <deliveryThreads> set to 1, and (2) later on with <deliveryThreads> set back to a higher value, just to help understand what happens and to test the fix. Does it make sense to do that way, or just do (1) or (2) right away? Secondly, is there anything in v2.1.1a5 I should know about before running it (remember I have users in production...)? Vincenzo -----Original Message----- From: Vincenzo Gianferrari Pini [mailto:[EMAIL PROTECTED]] Sent: gioved� 30 gennaio 2003 9.25 To: James Users List Subject: RE: CPU loop in "heavy" production environment Noel, I will do the following, in order to split the problems: 1) Run James today with the new <deliveryThreads> setting to 1, and see if the CPU problem disappears; 2) Run tomorrow with the James v2.1.1a5 test build, and see what happens to the exceptions; 3) Unless 1) is KO, I'll have to wait before moving to Sun JVM 1.4.1, as there is other stuff (Tomcat etc) running on the same machine. I'll let you know "on the fly" Thank you, Vincenzo -----Original Message----- From: Noel J. Bergman [mailto:[EMAIL PROTECTED]] Sent: gioved� 30 gennaio 2003 8.58 To: James Users List Subject: RE: CPU loop in "heavy" production environment I've heard of problems with Sun JVM 1.4.0 on SMP systems with linux. Supposedly, JVM 1.4.1 should address those. I can't speak for other operating systems. That might account for the CPU problem. Some of the exceptions, though, seem likely to be related to the defect I fixed earlier today, which was a synchronization error in the file system repository. I look forward to hearing from you. --- Noel -----Original Message----- From: Vincenzo Gianferrari Pini [mailto:[EMAIL PROTECTED]] Sent: Thursday, January 30, 2003 2:37 To: James Users List Subject: RE: CPU loop in "heavy" production environment I'm using Sun j2re1.4.0. Noel, as you said in another reply to this thread, the problem could be related to being on a dual CPU, also because I had changed the config entry <deliveryThreads> in the "remoteDelivery" mailet from 1 to 5 (without any particular reason, just to try/exploit multithreading under a dual CPU). So, before trying the James v2.1.1a5 test build, I will try today using <deliveryThreads> 1 </deliveryThreads>, and see what happens (already restarted James). I will let you know as soon as it doesn't work, or after 17.00 GMT if it does work. In the first case I will then try your fix, in the latter I will wait for your advise before trying the fix. In the meantime, thank you for the help. Vincenzo -----Original Message----- From: Noel J. Bergman [mailto:[EMAIL PROTECTED]] Sent: gioved� 30 gennaio 2003 5.42 To: James Users List Subject: RE: CPU loop in "heavy" production environment > RemoteDelivery.run(): java.util.ConcurrentModificationException I found what might have caused this, and would have cascaded to at least one of the other two problems. The fix is in the James v2.1.1a5 test build, which I will be uploading later tonight. Please test it, and let me know. --- Noel --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
