[ic] mod_interchange and Apache MaxClients

Ron Phipps rphipps at reliant-solutions.com
Thu Nov 17 11:04:16 EST 2005

> From: interchange-users-bounces at icdevgroup.org
> bounces at icdevgroup.org] On Behalf Of John1
> Sent: Thursday, November 17, 2005 2:47 AM
> On Wednesday, November 16, 2005 11:03 PM, kevin at cursor.biz wrote:
> > Here's another test to add to the list:
> >
> > When everything appears to pause, block port 80 at the firewall for
> > little while.  Remove the block and check whether the backlog has
> > cleared and IC is responding once again.  If this test succeeds then
> > the problem is flood/backlog related.  If not then, well, it was
> > a try.
> >
> In effect, I think I have verified that Apache doesn't start
> again even if port 80 were to be blocked.  Once MaxClients has been
> Apache stops servicing requests as it won't spawn any more processes
> for
> whatever reason non of the Apache processes ever die after this point,
> even
> if the server is left for many hours over night when I know the
traffic to
> my site is very low indeed.  Ron & Jeff - do you also see this effect?
I will take a look next time the event happens while I'm in the office.
What you describe seems to be what happened when MaxClients was reached
last time.  Lots of processes running, but none of them getting anything

I did notice some MySQL processes taking more time and processor usage
then normal last night so I decided to run an optimize on all of my
tables.  The site definitely sped up and the site did not need a restart
last night.  I'm not sure if they are connected or not, but perhaps John
and Jeff can run optimize on their tables and then see if the issue
continues to happen.


More information about the interchange-users mailing list