[Phpwm] Problems with live Silverstripe website (httpd spawing proceses, high CPU)

Pete Graham petegraham1 at gmail.com
Thu Aug 19 12:46:11 UTC 2010


Hi Adi,

Website is for INQ Mobile (http://inqmobile.com/) as mentioned before
company I work for (http://piratalondon.com/) didn't build the site
but we're maintaining it now and adding some new sections.

httpd.conf values are now:
KeepAlive On
MaxKeepAliveRequests 100
KeepAliveTimeout 15

<IfModule prefork.c>
StartServers 5
MinSpareServers 10
MaxSpareServers 30
ServerLimit 500
MaxClients 500
MaxRequestsPerChild  400
</IfModule>

<IfModule worker.c>
StartServers 20
MaxClients 500
MinSpareThreads     25
MaxSpareThreads     75
ThreadsPerChild     25
MaxRequestsPerChild  0
</IfModule>

Thanks everyone for your suggestions and help,
Pete

On 19 August 2010 13:25, Adi <aditya.r.d at gmail.com> wrote:
> So Pete, what are your new settings? If you don't mind telling which website
> are you talking about that someone is scraping 1000 pages per minute off.
>
> -
> Adi
> Same one from nzphplist
>
>
> On Thu, Aug 19, 2010 at 5:15 PM, Pete Graham <petegraham1 at gmail.com> wrote:
>>
>> Hi,
>>
>> I had my systems admin friend help me out on this one. He modified the
>> Apache configuration: tuned the prefork mpm. Re-activated keepalive,
>> with stricter settings. Also he noticed some IP in India was
>> requesting 1000 pages a minute sometimes so blacklisted the IP,
>> amongst other things.
>>
>> Site seems to behaving itself much more now, thankfully.
>>
>> Pete
>>
>> On 19 August 2010 12:33, David Goodwin <david at codepoets.co.uk> wrote:
>> > On Wed, 2010-08-18 at 17:09 +0100, Pete Graham wrote:
>> >> Hi everybody,
>> >>
>> >> I have a live SilverStripe site I've inherited that I'm experiencing
>> >> major problems with. After the site has been running for a while it
>> >> starts spawning multiple httpd processes, some taking up huge
>> >> percentages of the CPU (these can be seen using top). The site
>> >> progressively become more and more unresponsive returning 500 errors
>> >> for requests.
>> >
>> > Is the server running mod_php?
>> >
>> > Is the server set to kill it's children after N requests, or do children
>> > live forever? - is Apache's MaxRequestsPerChild 0 or something else?
>> >
>> > ( Perhaps it's just some sort of memory leak...? )
>> >
>> > I'd suggest you make sure error_log is enabled; and check apache's
>> > access logs.... perhaps it's possible to identify a request with
>> > something going wrong?
>> >
>> > I'd guess you could 'strace' the apache process that's misbehaving
>> > (assuming mod_php); but this might not work; or be at too low a level.
>> > I've never tried it myself.
>> >
>> >
>> > David.
>> >
>> >
>> >
>> > _______________________________________________
>> > Phpwm mailing list
>> > Website : http://www.phpwm.org
>> > Twitter : http://www.twitter.com/phpwm
>> > Facebook: http://www.facebook.com/group.php?gid=2361609907
>> >
>> > Post to list: Phpwm at mailman.lug.org.uk
>> > Archive etc : https://mailman.lug.org.uk/mailman/listinfo/phpwm
>> >
>>
>> _______________________________________________
>> Phpwm mailing list
>> Website : http://www.phpwm.org
>> Twitter : http://www.twitter.com/phpwm
>> Facebook: http://www.facebook.com/group.php?gid=2361609907
>>
>> Post to list: Phpwm at mailman.lug.org.uk
>> Archive etc : https://mailman.lug.org.uk/mailman/listinfo/phpwm
>
>
> _______________________________________________
> Phpwm mailing list
> Website : http://www.phpwm.org
> Twitter : http://www.twitter.com/phpwm
> Facebook: http://www.facebook.com/group.php?gid=2361609907
>
> Post to list: Phpwm at mailman.lug.org.uk
> Archive etc : https://mailman.lug.org.uk/mailman/listinfo/phpwm
>



More information about the Phpwm mailing list