Memory issues on Linux

HCL Notes/Domino 8.5 Forum (includes Notes Traveler)

Subject: Memory issues on Linux

Feedback Type: Problem

Product Area: Domino Server

Technical Area: Crash

Platform: Linux client - RHEL

Release: 8.5

Reproducible: Intermittent

Hello,
I'm running an 8.5 server on a vps Linux server, running CentOS 5.2 64 bits (technically equivalent to RHEL 5.2 64 bits). The box has 2GB RAM available. The server is quite unstable; it often crashes with the error printed below, but I have also seen the http and/or server process go to 100% CPU. I have tried several combinations of ConstrainedSHM and other memory related notes.ini settings, but haven't found the winning combination yet. Any recommendations as to how I should configure memory for this server?

I also wonder whether tunekrnl is able to set the right values in the kernel areas it's supposed to optimize; if I start the server manually, I get 'permission denied' errors there:

Error - can't open /proc/sys/fs/file-max.
errno: 13
Permission denied
Error - can't open /proc/sys/kernel/shmmax.
errno: 13
Permission denied
Error - can't open /proc/sys/kernel/sem.
errno: 13
Permission denied
Error - can't open /proc/sys/net/ipv4/tcp_fin_timeout.
errno: 13
Permission denied
Error - can't open /proc/sys/net/ipv4/tcp_max_syn_backlog.
errno: 13
Permission denied
Error - can't open /proc/sys/net/ipv4/tcp_tw_reuse.
errno: 13
Permission denied
Error - can't open /proc/sys/net/ipv4/ip_local_port_range.
errno: 13
Permission denied

Can anyone tell me what values should be in these files, so that I can compare them?

This is an example of the error sometimes observed on the system:

** Error attaching to shared memory [loc=MapSharedRegion,error code=12,ptr=ffffffff] -- check Notes configuration variables (ConstrainedSHMSizeMB, PercentAvailSysResources).

Thread=[03976:00002--173762336]
Stack base=0xFFCC166C, Stack size = -1172 bytes
PANIC: Cannot attach to shared memory region, due to insufficient access (probably owned by another user or group)
Stack base = 0xffcc166c, Stack size = 332 bytes
Fatal Error signal = 0x0000000b PID/TID = 3976/-173762336
4/8/2009 2:19:13 Running NSD
4/8/2009 2:19:13 Fault cleanup is in progress
4/8/2009 2:19:13 Too many crashes, restart disabled
4/8/2009 2:19:13 Terminating tasks
4/8/2009 2:19:27 Freeing resources
4/8/2009 2:19:27 Termination completed

Thanks in advance, kind regards,
Gerco

Feedback number WEBB7QW9JS created by ~Dan Umjipylit on 04/08/2009

Status: Open
Comments: