RE: Crash on 2.5PRE7-CVS

> Hello,
>
> We're seeing various problems with that release under load as well. It
> appears once the squid process grows to around 100MB (which at roughly
> 300 req/second is about an hour) it becomes useless, almost
> totally stops
> serving connections. Systems run 512MB ram.

Diffrent problems. My squid was 1.3Gb in size (1 Gb memcache) when
it died.

> I'm not really sure what the right combo of servers +
> hardware is gonna be
> to route 100mBit/sec of traffic through squid.

I'm using Alteon load balancers for load-balancing,
and cachedigests+icp for inter-cache communications,
NTLM authentication via SMB helper and extensive ACLs.
In conditions similar to my setup, I think that for that
kind of traffic you'd need 5 to 6 dual-proc 1GHz servers,
each with 2 instances of Squid running. 8-10 servers if
you want a single instance per server.

-- 
	/kinkie