ouch

[Chris]

my server was down, I managed to log in for a brief few seconds and get this readout:



top - 13:01:05 up 119 days, 17:06, 1 user, load average: 3513.96, 3529.80, 3531.06
Tasks: 9681 total, 13 running, 9667 sleeping, 1 stopped, 0 zombie
Cpu(s): 0.7% us, 99.0% sy, 0.0% ni, 0.0% id, 0.2% wa, 0.0% hi, 0.0% si
Mem: 10038460k total, 10019660k used, 18800k free, 1840k buffers
Swap: 2048276k total, 2048152k used, 124k free, 23768k cached


lots of exim and httpd processes as I can tell.

[rpanella]

Wow! Any idea what caused that?
________
Dodge ram (disambiguation) specifications (http://www.dodge-wiki.com/wiki/Dodge_Ram_(disambiguation))

[Chris]

I'm not entirely sure.

[Micha Dugga]

I am amazed that you were even able to log in. I have a remote IP-based KVM on our rack of servers which gives me local keyboard access, and if any of my boxes load to over 100 or so, I can't log in even locally!

Nearly 10,000 processes. Wow. Usually, Apache's max_servers config limit keeps the server from being taken down like that. I'm not as familiar with exim, but I know sendmail will start rejecting mail by default at a certain load level also.

Weird. Well, I guess things are OK now?