Re: [squid-users] high cpu usage

A small number of url_regex acls shouldn't cause a huge burden on your
box. It is primarily a concern when there are many (20+ probably).

Pat Lendon wrote:
> Thank you for responding.
>
> I have a couple of url_regex acls:
>
> acl sysadm url_regex ~http://xx.xx.x.xxx
> always direct allow sysadm
> acl worm url_regex ~http:/xx.xx.x.xxx
> http_access deny worm
>
> How do I form the dstdomain acls to replace above two url_regex acl?
> acl sysadm dstdomain sysadm.microcenter.com
> acl worm dstdomain sysadm.microcenter.com
>
> Again - thank you!
>
>
> Peter Arnold wrote:
>
>> Just as a matter of interest, do you have any/many access statements
>> using regex
>> ACLs?
>>
>> I had a recent experience (2.5p3) where the (subjective) over use od
>> url_regex
>> REALLY bogged down the server. Changed these to dstdomain and all was
>> fine.
>>
>> Cheers
>> Peter Arnold
>>
>> Joe Cooper (joe@swelltech.com) wrote*:
>>
>>> This isn't expected behavior for Squid, so tell us what configure
>>> options you used...
>>>
>>> Pat Lendon wrote:
>>>
>>>> Hello All,
>>>>
>>>> Randomly, the cpu usage on system shows squid at 99%. The cache log
>>>> does not contain any information at the time cpu was pegged. I copied
>>>> the cache manager log.
>>>> We are running Squid 2.4.STABLE4 on RedHat Linux. This is a brand new
>>>> system with two AMD Athlon(tm)MP 1800+ processors and 1gig memory. Any
>>>> ideas on cause of cpu at 99%?
>>>> Is the cache causing high cpu usage. I noticed the page faults with
>>>> physical i/o
>>>> at 402. Is this a concern?
>>>>
>>>
>>
>>
>
>
>
>
>

-- 
Joe Cooper <joe@swelltech.com>
http://www.swelltech.com
Web Caching Appliances and Support