On Tue, 4 Nov 2003, Netguy wrote:
> i get now around 24% bytehitratio other compititors in my country is getting
> around 40% with 1TB cache using cacheflow
Have you done any refresh_pattern tunings of your Squid? This can make a
big difference.
What is the LRU reference age of your cache_dir?
On Tue, 4 Nov 2003, Netguy wrote:
> i think freebsd support upto 2GB for single process right?
Not a FreeBSD guy, but it quite likely does. But it is also quite likely
you need to do some kernel tuning to allow such large processes (but maybe
not, as I said I am not a FreeBSD guy. All I know is
> i get now around 24% bytehitratio other compititors
> in my country is getting around 40% with 1TB cache
> using cacheflow
Last week, we achieved a 50% HTTP hit ratio using under 1 GB of cache.
Have you tried adjusting the refresh_pattern settings in Squid? This
can help raise your byte hit rati
?
Netguy
- Original Message -
From: "Henrik Nordstrom" <[EMAIL PROTECTED]>
To: "Netguy" <[EMAIL PROTECTED]>
Cc: <[EMAIL PROTECTED]>
Sent: Tuesday, November 04, 2003 12:06 PM
Subject: Re: [squid-users] large storage
> On Tue, 4 Nov 2003, Netguy wrote:
>
On Tue, 4 Nov 2003, Netguy wrote:
> i am planning to buy a machine with large storage (around 1 or 2
> tera-byte) and running it as squid proxy.
Do you really need that large storage in a single Squid proxy? There is
very little return when growing the cache beyond 1 weeks worth of content.
> i
Hi,
i am planning to buy a machine with large storage (around 1 or 2 tera-byte)
and running it as squid proxy. i have now a machine with 125GB cache with
cache_mem set to 256 and the squid process grows up to 1.3GB. the problem is
in my openion in the memory because the squid process will grow ab
