Re: [squid-users] High cpu load with squid

On 18/11/10 08:14, Michał Prokopiuk wrote:
> Dnia Poniedziałek, 01.11.2010 o 2:17 Amos Jeffries napisał:
>> On Sun, 31 Oct 2010 20:45:33 +0100, Michał Prokopiuk
>> <michal_at_linuxstuff.pl>
>> wrote:
>>> Hi.
>>> For the last few days I have had troubles with my squid (3.0.24) - it
>> gets
>>> a lot of cpu, and
>>> response time is very long, about 4000 ms. I didn't see any information
>>> in cache.log, so I tried to upgrade to latest stable release - it
>>> doesn't change anything. I have read a lot about high cpu usage when
>> squid
>>> rebuild disc cache, so I try change cache_dir type from ufs to aufs or
>>> coss (./configure said about remove coss from squid 3.x), but the
>> problem
>>> still exists.
>>>
>>> What am I doing wrong? My squid has been working perfectly for two years
>>> :(. Here's my config.log and squid.conf:
>>>
>>> ./configure --prefix=/usr/local/squid-3.1.8 --enable-useragent-log
>>> --enable-referer-log --enable-default-err-language=Polish
>>> --enable-err-languages=Polish --enable-linux-netfilter --enable-ssl
>>> --with-large-files --with-filedescriptors=16384 -enable-snmp
>>> --with-maxfd=16384 --with-pthreads --enable-async-io --disable-dlmalloc
>>> --with-aio --enable-epoll
>>>
>>> http_port 192.168.1.1:8080 transparent
>>> pid_filename /var/run/squid/squid.pid
>>> acl sloneczko src 192.168.0.0/23
>>> http_access allow sloneczko
>>> error_default_language pl-pl
>>>
>>> redirect_program /usr/local/bin/redirector.pl
>>> redirect_children 80
>>
>> Replace those with:
>> url_write_program /usr/local/bin/redirector.pl
>> url_rewrite_children 80
>>
>>>
>>> store_avg_object_size 8 kB
>>> minimum_object_size 1 KB
>>> maximum_object_size 100 MB
>>> maximum_object_size_in_memory 1 MB
>>> cache_mem 1000 MB
>>>
>>> cache_swap_low 80%
>>> cache_swap_high 100%
>>
>> cache_swap_high should be something less than 100% (99% may be better).
>> Squid will ONLY begin the aggressive space clearing when cache_swap_high
>> threshold is passed. So with 100% this may cause active connections to be
>> stopped while 20% of the cache is discarded.
>>
>>> # previous cache
>>> #cache_dir ufs /var/spool/squid-cache/cache 30000 12 256
>>> cache_dir aufs /var/spool/squid-cache/cache 30000 60 100
>>>
>>> dns_nameservers 192.168.1.100 192.168.1.200
>>> ipcache_size 8192
>>> fqdncache_size 1024
>>> positive_dns_ttl 2 hours
>>> negative_dns_ttl 1 minutes
>>> ipcache_low 90
>>> ipcache_high 95
>>>
>>> emulate_httpd_log on
>>> access_log /var/log/squid/access.log
>>
>> Remove emulate... and change access_log to:
>> access_log /var/log/squid/access.log common
>>
>>> cache_log /var/log/squid/cache.log
>>> cache_store_log /dev/null
>>
>> Set this to "cache_store_log none".
>>
>>>
>>>
>>> Rest of squid.conf are acls. I have about 30 - 40 mbps of traffic. On
>>
>> You have "http_access allow sloneczko" at the top of this config. so its
>> possible your other ACL are not working.
>>
>>> board
>>> are core2duo 1.8 ghz 4 gb ram on intel chipset, 2x SATA on soft raid
>> (for
>>> cache - md0).
>>
>> by "soft raid" you mean *software* raid? That is a disk IO killer for
>> Squid.
>>
>>>
>>> Any ideas? When I run -k debug I can't do anything - load has increased,
>>
>
> I set all canges, unfortunately, squid work great for fiew days, and today problem back again. Any other ideas? I didn't see any other solution except remove squid :(
>

Okay. The obvious config reasons for slowness are gone. So the next
thing is to find out exactly what Squid is doing.

* you can check with strace to see whats the thing taking out the CPU.

  Has the box started using swap memory space now? That can drastically
slow downs Squids index lookups or buffering of information relayed.

  Is RAID still running trying to duplicate the random cache IO actions?

  Have you noticed if the problem appears and lasts for around 2 hours,
then goes away? that would indicate your DNS TTL overrides forcing squid
to try the wrong IPs for some site. CDN hosted sites (eg akamai and
youtube) can change their IPs with no notice when there are routing
problems to work around.

Amos

-- 
Please be using
   Current Stable Squid 2.7.STABLE9 or 3.1.9
   Beta testers wanted for 3.2.0.3