cache size questions and dns problems

From: Steve Pfister <srp336@dont-contact.us>
Date: Tue, 14 Dec 1999 15:11:37 -0600

Hello--

We are running Squid 2.2stable4 under Solaris 2.6 on a Sun Ultra5. I've got
a few questions on the optimal set up for our users.

- We occasionally get a DNS error when requesting pages. It says "cannot
find server or DNS error." Trying again immediately will give us the
requested page. We've got dns_children set at the default of 5...should we
raise this? Is this a Squid configuration problem, or something else?

- We have about 150 users for this proxy. The cache is currently at 100mb,
with L1=7 and L2=64. After several days of running, our storage LRU
expiration age is about 0.85 days, and the mean object size is about 8kb.
How do we determine the best choice for cache size? Should we aim for a
certain value for the LRU expiration age, say 3-5 days?

- How do we rotate logs? Both "squid -k rotate" and "kill -USR1 <pid>"
don't seem to do anything.

- How do we log domain names in access.log? The option log_fqdn is left at
its default of on (log_fqdn is not commented out), and nslookup returns the
right answers, but we still get ip addresses in the log.

- How is the cache directory indexed? I thought this might be in the FAQ,
but couldn't see it.

- Are there any log analyzers that are popular and work well with Squid logs?

Thanks!

--Steve Pfister
Received on Tue Dec 14 1999 - 13:17:56 MST

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:49:54 MST