Hello all,
Recently we setup a Debian Linux box solely to run Squid (version
1.1.21) as a caching proxy server. All seemed well to get it up and
running and get a squid.conf file that seems fairly appropriate for the
time being. We now would like to tweak things a bit to get this ship
running tight. I would extremely appreciate a few answers to my
questions below, if anyone has the time and patience:
(1) Currently out IPCache is set to 1024 (the default). Anyone have any
suggestions on what might be a better number, or perhaps how to figure
out what would work better for us.
(2) Using cachmgr.cgi I notice that after a few days everything seemed
to even out, and memory usage looks to be good. However, the Avg.
Name-Address lookup time suddenly went negative (right now it sits at
-3.399000 seconds). How can the lookup time be negative?
(3) Since this machine has *plenty* of RAM (512 MB) and has no other
real purpose, we pulled out anything that looked like it might hinder
performance. This included turning memory_pools on. Any suggestions on
this kind of setup?
(4) I notice that the Maximum Resident Size is at 0 KB. There is also
the line "Page faults with physical i/o: 8204" underneath it. The FAQ
discusses that the RSS should be close to the Total Space in Arena.
Obviously there is a huge gap. The page fault # that is referenced in
the FAQ (4897) is different than the number we are seeing. What gives?
How can we get this RSS # set correctly?
I sincerely appreciate anyone who can give some suggestions and/or
answers. Thanks in advance!
Sincerely,
Marcus Hecht
-- ======================================================= Marcus Hecht PCs Made Easy marcus@pcez.com Internet Services Support Specialist http://www.pcez.com =======================================================Received on Mon Aug 03 1998 - 14:31:00 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:41:26 MST