On 10/23/2013 11:44 AM, Amos Jeffries wrote:
Search in cachemgr report "events" and see what will run on the same
time when the outage appears. That will give you a good idea to look.
You might also grab a copy of "info" report before and after the outage
time. Compare the cache object count, and avg stored object size. If the
avg size is small your Squid will be doing a large amount of work
clearing up old cache content out of the 300GB+ storage you have.
Amos
Which means it's caching too much??
Dont we have a calculator that tells you how much you cache_dir size to
use??
By the way the benefit of couple small disks is that you have instead of
3Gbps bandwidth to share all the I\O to the disk is that you have 3Gbps
per DISK for both I\O traffic.
which means that it can happen that squid will try to access 1000
objects at the same time from the same disk But I assume it's not
happpning every second.
Eliezer
