On 30/10/2015 05:24, Patrick Blair - Peapod wrote:
Hi Eliezer,
Thanks for your response.
I have set up a VM to test out configurations in the same data center and
address space as the problematic one.
What I haven't done is test it by rebuilding the squid configuration from
the defaults up and trying to use the same IP, that will probably be what
I'll try tomorrow.
First try to test it with default squid.conf and only adding to the
localnet the local networks address space\subnet on the current proxy.
Also, thanks for the tip on the CARP example. I was trying to find a
configuration that took advantage of SMP, but I see how that complicates
things further.
SMP can work without any relation to the CARP example.
It is limited to rock cache_dir and ram cache.
I would start with a RAM cache only and a number of workers.
The number of workers you will be required to use depends on your
network load.
You can try with 2 workers and use the squid cache manager interface to
see if you have high requests per second rate per worker.
If you need some hints on that I will try to find the exact values in
the cache manager interface.
Thanks so much for the suggestions! I'll update this thread further if
things start working a bit better.
And thank you again for packaging newer squid versions for CentOS!
Appreciated, I am trying my best with my limited resources.
Eliezer
_______________________________________________
squid-users mailing list
squid-users@xxxxxxxxxxxxxxxxxxxxx
http://lists.squid-cache.org/listinfo/squid-users
