Is this machine too slow? Kernel cannot handle page request error.
I have this machine set up as a Postfix/Spamassassin mail relay server for our LAN (125 users). When it works, it kicks serious *ss. The problem is it crashes (completely) every couple of days under load. The error message in /var/log/messages reads
[Kernel] Unable to handle page request. and then a whole bunch of gibberish. The machine specs are as follows p-II 266, 256mb ram, 10Gb hdd, two nics (one 3com, one lynksys isa), and a slim installation of Redhat 9 running only Postfix, Spamassassin, and Iptables as services. Is this indicative of a memory issue? Or is the machine just too old to serve in this capacity? :scratch: |
How large is your swap partition? It almost sounds like the kernel was trying to read a page from memory, but an error occured in the read/write process. I'm not sure what kind of requirements the apps that your running need, but I would think that 256 MB would be a wee bit on the small side for any app that is serving 125 users. Any chance you can post what the bunch of gibberish is under the error message?
slight |
My swap partition (created automatically) is 512mb, approximatly twice the size of the physical ram. It does scream memory problem.
|
Try running Memtest86 on it and maybe you'll find a memory issue. Another diagnostic tool I would suggest is Gkrellm. Even if you are not running X on the server box I think you can run Gkrellm on it and then connect to it using another machine running X.... at least this would give you a real time look at memory usage and swap usage.
slight |
I will give that a shot right now. Thanks a lot for the advice and the link. Hope this works. Spamassassin is really cool.
|
All times are GMT -5. The time now is 05:48 AM. |