The Community Forums

Interact with an entire community of cPanel & WHM users!
  1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Apache Goes Out To Lunch for 1 to 2 minutes at a time

Discussion in 'EasyApache' started by gator, Mar 28, 2005.

  1. gator

    gator Active Member

    Joined:
    Feb 27, 2003
    Messages:
    37
    Likes Received:
    0
    Trophy Points:
    6
    OK, we are stumped on this one and would welcome any ideas you all may have....

    We have a CPanel server that is running fine in every respect except that Apache will suddenly stop responding for 1 to 2 minutes at a time. Then return to normal and work fine again. During these brief outages access to the server in every other way works flawlessly and lightning fast (ping, mail, traceroute, ssh, work fine). Yet no response comes from Apache. The request times out in the user's browser. We can go several hours without a problem and then hit a window where it seems to occur frequently and then settle down for a while. It does not appear to coincide with server load though.

    We have combed through every log file we can think of and found no errors that would explain this behavior. We increased MaxClients but our monitoring program indicates we are averaging about half the current max clients count (it checks every 15 minutes). I also thought CPanel emails us when MaxClients is reached - that hasn't happened.

    When apache wakes up again it is fast and no unusual delays. Server load and memory are all acceptable and look great.

    The data center has looked at the network and believes everything is fine. By why would it just drop port 80/443 traffic and nothing else? We upgraded to the latest copy of APF firewall on our Fedora 2 box and no real change. We log dropped packets in APF and see nothing dropped for apache for the given IP's.

    Increasing MaxClients seemed to reduce the frequency slightly (we went from 150 to 300) but it does not appear to be the issue.

    Any idea of what we should grep for to see if we can identify why this is happening? If anyone has any ideas they are most appreciated. Thanks!
     
  2. Blue|Fusion

    Blue|Fusion Well-Known Member

    Joined:
    Sep 12, 2004
    Messages:
    378
    Likes Received:
    0
    Trophy Points:
    16
    Location:
    Cleveland, Ohio
    When Apache becomes unresponsive, login to SSH and see what the httpd processes are doing, if anything. If they all become defunct, sounds like a script is crashing Apache. If they aren't there, then something is definately crashing Apache. If niether of the two problems are present, try using wget to get a webpage on your server, and see if that times out, too. If that times out, then you can be 100% certain it isn't the network. I highly doubt APF would be at fault. I've never had a problem with it and there's no reason for it to start dropping packets on those 2 ports all of a sudden.
     
  3. Kerstin

    Kerstin Well-Known Member

    Joined:
    Apr 9, 2005
    Messages:
    136
    Likes Received:
    0
    Trophy Points:
    16
    Location:
    Berlin
    Existing apache clustering for secure weblogin ? :mad:

    When ip 2xx.xxx.xxx.xxx not acessable :mad: ;)
    and rerouted to the cluster ip 6x.x.xxx.xxx.xxx ;).

    Forbidden error by not scure ip-login ;) ?

    :confused:
     
  4. webignition

    webignition Well-Known Member

    Joined:
    Jan 22, 2005
    Messages:
    1,880
    Likes Received:
    0
    Trophy Points:
    36
    Kerstin, would you care to clarify things somewhat by explaining what you are trying to do/achieve, what the actual result is and what the desired result is (i.e. what should happen compared to what is actually happening). Your query is a little vague.
     
Loading...

Share This Page