On 18/08/2012 12:09 p.m., Ricardo Rios - Shorewall List wrote:
El 2012-08-17 18:01, Ricardo Rios escribió:
On 17/08/12 14:43, Kinkie wrote:
On Fri, Aug 17, 2012 at 5:37 PM, Ricardo Rios - Shorewall List
<shorewall@xxxxxxxxxxxxxxx [1]> wrote:
I am trying squid 3.2.1 in a opensuse 11.4 Box, with 4 workers and 2
rock cache_dirs with 500 clients browsing, all looks perfect at
start, but before 5 or 10 mins, i see the traffic going down and up,
when i check top, and this happend, i see the squid process going
defunct, after anothers 5 or 10 mins, happend again, and again, and
again.
Anything in cache.log?
Nop, witch is weird, i gonna try run a ./squid -X and try to see what
error i get.
I only see this on cache.log :
http://pastebin.com/XJeUmPES
Where is all that javascript coming from? it is not related to any of
the logging lines around it in your paste.
and after a few mins i see this :
http://pastebin.com/pe3QXsGj
Squid loading its error page templates. With a warning that you have
none installed for the cache mgr front page (new cachemgr interface
feature in 3.2).
I try running just without "workers 4", i try without using any
cache_dir, and same results. Guess i should go back to 3.1.19 witch i
got working well. I give a try to 3.2.1 since rock cache_dir and SMP
looks good.
You should be able to use 3.2.1 just fine with only the same features
you were using under 3.1. It *is* measurably better in a number of basic
functionality. By the sounds of it I think you may be able to use just
rock store without workers.
If you could assist with figuring out this weird problem, even better.
Amos
