[Imc-tech] Website Sloooow

[Mike Lehman]

Just a wild ass guess from someone who knows nothing about Apache...

But the problems we have been having with this seem to correlate to some 
degree with the increased number of article delete purges that I've had 
to do to deal with Jack Ryan. Could there be causation to go with 
correlation, re Dada and Apache?
Mike Lehman

Arun Bhalla wrote:

> Everytime I flip through the error.log, I see bunches of "Segmentation fault"
> and maybe some other errors.  I don't know if it strongly correlates with
> Apache spinning or being unresponsive.  Either way, it seems like a problem.
> I usually don't see that in Apache error logs.  But it would be great to
> be able to correlate the PIDs in the error logs with the requests.
> 
> Unfortunately I didn't take note whether it was the master Apache process
> or the child servers that was the runaway.
> 
> Arun
> 
> "Zachary C. Miller" writes:
> 
>>Ok this runaway apache problem is driving me nuts! 
>>
>>We need to get to the bottom of what is causing it. I swear, I have to
>>restart apache like once a day. It seems like rebooting staves off the
>>problem for a day or two and then it comes back. I wonder if it is
>>related to going deep into swap.
>>
>>No matter what I set the various timeouts in httpd.conf to doesn't
>>seem to help at all. 
>>
>>The stuck processes are never on any kind of exotic or difficult PHP
>>request, they're usually for something very mundane and random.
>>
>>Upgrading apache hasn't helped, apparently. 
>>
>>Any ideas on how to proceed with debugging this folks? 
>>
>>Arun Bhalla wrote:
>>
>>>There was a runaway Apache process with 50+% CPU.  Restarting Apache
>>>to get rid of that...
>>>
>>>Arun
>>>
>>>Arun Bhalla writes:
>>>
>>>>I checked it out.  It looks like the server is currently processing
>>>>(esp. virus and spam scanning) a bunch of emails that are being
>>>>delivered locally.  Most of the CPU is being spent on that, and there's
>>>>little free RAM, so Apache is getting swapped out.
>>>>
>>>>I'm going to let those emails process.
>>>>
>>>>Arun
>>>>
>>>>Mike Lehman writes:
>>>>
>>>>>I just wanted to note, that even though it doesn't seem to have gone 
>>>>>down, the website has been extremely slow all morning. This may have 
>>>>>also been the case overnight, as we had a number of duplicate posts 
>>>>>made, which could indicate that people were having trouble posting and 
>>>>>may have tried posting again, when their first posts were simply taking
>>
>> 
>>
>>>>>a very long time to post.
>>>>>Mike Lehman