On 31/07/2015 8:05 a.m., Josip Makarevic wrote: > Hi, > > I have a problem with squid setup (squid version 3.5.6, built from source, > centos 6.6) > I've tried 2 options: > 1. SMP > 2. NON-SMP > > I've decided to stick with custom build non-smp version and the thing is: > - i don't need cache - any kind of it cache_mem 0 cache deny all That is it. All other caches used by Squid *are* mandatory for good performance. And are only used anyway when the component that needs them is actively used. > - I have DNS cache just for that > - squid has to listen on 1024 ports on 23 instances. > each instance listens on set of ports and each port has different outgoing > ip address. And how many NIC do you have that spread over? > > The thing is this: > It's alll good until we hit it with more than 150mbits then... > > (output from perf top) > 84.57% [kernel] [k] osq_lock > 4.62% [kernel] [k] mutex_spin_on_owner > 1.41% [kernel] [k] memcpy > 0.79% [kernel] [k] inet_dump_ifaddr > 0.62% [kernel] [k] memset > > 21:53:39 up 7 days, 10:38, 1 user, load average: 24.01, 23.84, 23.33 > (yes, we have 24 cores) > Same behavior is with SMP and NON-SMP setup (SMP setup is all in one file > with workers 23 option but then I have to use rock cache) > > so, my question is....what...how to optimize this.....whatever....I'm stuck > for days, I've tried many sysctl options but none of them works. > Any help, info, something else? None of those are Squid functionality. If you want help optimizing your config and are willing to post it to the list I am happy to do a quick audit and point out any problem areas for you. But tuning the internal locking code of the kernel is way off topic. Amos _______________________________________________ squid-users mailing list squid-users@xxxxxxxxxxxxxxxxxxxxx http://lists.squid-cache.org/listinfo/squid-users
