My disc IO is suddenly 10x that what is averages for the past 6 hours.

How do I best figure out who is causing this?

Running centos with mysql/lighttpd, traffic numbers are about the same as yesterday.

Hi, you can use iotop for that (at least in Debian) maybe there's something similar in Centos.

Another interesting tool is iostat (is in the sysstat package in Debian but I don't know in Centos).

thanks, those tools were useful.

Looks like its mysql:

0 0 23 76 0 0|mysqld 0.2| 0 36M

Writing out 36M is not good I think. I guess I need to optimize my mysql?

I was running into the same problem. Disk I/O suddenly jumped to 5X normal overnight. I'm running the Postgres database. I did a "vacuum analyze" to see if that would help, and disk I/O went back down to normal levels. I thought I had that in my daily cron job, but don't see it now.

Dave