I've done a lot of reading, and I can't seem to figure out what's going on. OS FreeBSD 4.10-RELEASE-p2. Squid is Version 2.5.STABLE6. There are 4 15k 18GB drives being used for the cache. It has a 3.0 GHz HT Intel processor, and 2 GB of RAM.

This install has been running since Monday. I noticed on Monday that squid kept dieing due to the default 512MB data seg siz. Besides that, squid was running flawlessly.

Last night I added the kernel options to allow max data seg size of 1.5 MB. Everything was running smooth until this morning around 10:00a.m. At this time, the CPU pegged out at 100%, so I restarted squid. I then watched squid climb back up to 100%. Every once in a while, it will start to climb back down, but then it will jump right backup. I can reproduce this behavior with just a single machine accessing the cache.

The cache is being run transparently, using IPFW to redirect.

Relevant config lines:

cache_mem 512 MB

cache_dir diskd /usr/local/squid/cache/1 10000 16 256 Q1=72 Q2=64
cache_dir diskd /usr/local/squid/cache/2 10000 16 256 Q1=72 Q2=64
cache_dir diskd /usr/local/squid/cache/3 10000 16 256 Q1=72 Q2=64
cache_dir diskd /usr/local/squid/cache/4 10000 16 256 Q1=72 Q2=64

cache_replacement_policy heap GDSF

Any help what-so-ever would be greatly appreciated!!

Thanks,
Russ
---
[This E-mail scanned for viruses by Declude Virus]



Reply via email to