Hi All, We have installed RH 9.0 and upgrade the initially to kernel 2.4.21, and then kernel 2.4.22. The bizzare out of memory problem persists 4-5 times last week. Machine runs out of memory and swap slowly and we couldn't identify why the system behaves like this, nothing strange in the process table. We are running Bind 9.2, Apache 1.3, RAdius server 0.18.1, MySQL 3.23.57 and postfix. Primary busy services on this server are only due to Apache and radius. Main database is as big as 30 Mb and request per second is no way more then 10-50. We suspect from the C libs on the system, gcc version 3.2.2 20030222 (Red Hat Linux 3.2.2-5) where malloc or realloc has some sort of stability issue, because we can compile radius on gcc 2.96 but not with 3.2.x. Every 2 days, aroun 08:30 system hangs and run out of memory. (No scheduled cron job is running at that time!)\ A part from syslog Aug 28 08:23:47 kernel: Out of Memory: Killed process 1020 (named). Aug 28 08:23:47 kernel: Out of Memory: Killed process 1027 (named). Aug 28 08:23:47 kernel: Out of Memory: Killed process 1028 (named). Aug 28 08:23:47 kernel: Out of Memory: Killed process 1029 (named). Aug 28 08:23:47 kernel: Out of Memory: Killed process 1030 (named). Aug 28 08:23:52 kernel: Out of Memory: Killed process 9881 (httpd). Aug 28 08:23:58 kernel: Out of Memory: Killed process 4920 (httpd). #> sar -r 07:10:00 AM kbmemfree kbmemused %memused kbmemshrd kbbuffers kbcached kbswpfree kbswpused %swpused 07:20:00 AM 301288 731828 70.84 0 132916 263936 987604 384 0.04 07:30:00 AM 300796 732320 70.88 0 133048 263980 987604 384 0.04 07:40:00 AM 300560 732556 70.91 0 133168 264324 987604 384 0.04 07:50:00 AM 300460 732656 70.92 0 133360 264428 987604 384 0.04 08:00:00 AM 300212 732904 70.94 0 133544 264648 987604 384 0.04 08:10:00 AM 299432 733684 71.02 0 133716 264752 987604 384 0.04 08:20:00 AM 300460 732656 70.92 0 133856 264824 987604 384 0.04 08:32:22 AM 5312 1027804 99.49 0 8812 41752 0 987988 100.00 09:20:01 AM 922020 111096 10.75 0 22076 43248 987988 0 0.00 09:30:00 AM 859704 173412 16.79 0 29160 95804 987988 0 0.00 09:40:00 AM 841696 191420 18.53 0 35528 101460 987988 0 0.00 Anybody could suggest any sort of diagnosis or suggestion why this might happend? Named or Httpd memory leakage? Radius or any other combination. By the way we are running Xeon proc on Intel Server Board SE750 1 BR2. Many Thanx for the comments for this bothering problem. Mehmet -- Shrike-list mailing list Shrike-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/shrike-list
