From: Masood R. <ma...@in...> - 2010-03-27 07:42:26
|
Didn't get any responses to my message below so I'm resending. Also I have 0 set for session limits but see the following entry in my logs: Mar-26-10 21:03:05 [Main_Thread] Warning : SMTP-session-limit 999 is reached - SMTP listeners will be switched off More possible problems. After disabling the regex assp does not die on 2 servers, however sometimes on both of these servers assp stops accepting new requests. So I looked at the maillog file during one of these problems and found the following: Mar-24-10 13:50:49 [Worker_12] Error: Worker_12 accept to client failed IO::Socket::INET=GLOB(0x118cc5c8) (timeout: 2 s) : Too many open files Mar-24-10 13:50:49 [Worker_12] Error: Worker_12 accept to client failed IO::Socket::INET=GLOB(0x11907ee0) (timeout: 2 s) : Too many open files Mar-24-10 13:50:49 [Worker_12] Error: Worker_12 accept to client failed IO::Socket::INET=GLOB(0x9b0fd500) (timeout: 2 s) : Too many open files Mar-24-10 13:50:49 [Worker_12] Error: Worker_12 accept to client failed IO::Socket::INET=GLOB(0x936fd6f0) (timeout: 2 s) : Too many open files Mar-24-10 13:50:49 [Worker_12] Error: Worker_12 accept to client failed IO::Socket::INET=GLOB(0x9342ec98) (timeout: 2 s) : Too many open files Mar-24-10 13:50:49 [Worker_12] Error: Worker_12 accept to client failed IO::Socket::INET=GLOB(0x936785f0) (timeout: 2 s) : Too many open files After that the server stopped responding to requests on port 25 but assp was still running. I find this error interesting because I had already increased the number of open files allowed for the assp user the server runs as several months ago to 65500 in the /etc/security/limits.conf file. I verified these changes were valid by doing 'ulimit -a' as the assp user. I find it hard to believe assp has 65,500 files open??? After all I am using the mysql db and berkeley db for most everything now. So how come assp has so many files open? Is it possible that assp does not close open files correctly and as it runs throughout time this number just keeps increasing? In addition to finding a resolution on this it would be nice if assp automatically restarted itself in this case (which it did not do). Because the restart would have fixed the issue. Issue #2: I have assp restart itself automatically in a cron job, it usually works fine. However this time it looked like there was a problem retrieving the griplist? See the error below. As a result ASSP never restarted correctly and just sat their not processing any incoming requests until I restarted it again. - BDB:Permission denied Mar-25-10 00:00:16 [Worker_10] BerkeleyDB-CRT-ENV-ERROR Griplist: BDB:Permission denied Mar-25-10 00:00:16 [Worker_10] BerkeleyDB-ENV-ERROR Griplist: Can't call method "set_timeout" on an undefined value at (eval 1894) line 80. - BDB:Permission denied Mar-25-10 00:00:17 [Worker_11] BerkeleyDB-CRT-ENV-ERROR Griplist: BDB:Permission denied Mar-25-10 00:00:17 [Worker_11] BerkeleyDB-ENV-ERROR Griplist: Can't call method "set_timeout" on an undefined value at (eval 1894) line 80. - BDB:Permission denied Mar-25-10 00:00:18 [Worker_12] BerkeleyDB-CRT-ENV-ERROR Griplist: BDB:Permission denied Mar-25-10 00:00:18 [Worker_12] BerkeleyDB-ENV-ERROR Griplist: Can't call method "set_timeout" on an undefined value at (eval 1894) line 80. - BDB:Permission denied Mar-25-10 00:00:18 [Worker_13] BerkeleyDB-CRT-ENV-ERROR Griplist: BDB:Permission denied Mar-25-10 00:00:18 [Worker_13] BerkeleyDB-ENV-ERROR Griplist: Can't call method "set_timeout" on an undefined value at (eval 1894) line 80. - BDB:Permission denied Mar-25-10 00:00:18 [Worker_15] BerkeleyDB-CRT-ENV-ERROR Griplist: BDB:Permission denied Mar-25-10 00:00:18 [Worker_15] BerkeleyDB-ENV-ERROR Griplist: Can't call method "set_timeout" on an undefined value at (eval 1894) line 80. - BDB:Permission denied Mar-25-10 00:00:18 [Worker_14] BerkeleyDB-CRT-ENV-ERROR Griplist: BDB:Permission denied Mar-25-10 00:00:18 [Worker_14] BerkeleyDB-ENV-ERROR Griplist: Can't call method "set_timeout" on an undefined value at (eval 1894) line 80. |