MaxClients constantly reached.

[roycrom]

Hi all. I know this is a long shot but here goes...

I have Apache 1.3.28 running on solaris 8. I have php-4.3.2, mysql-4.0.14, postgresql-4.3.2, gd-2.0.15, jpeg-6b, imagick-5.5.7, libpng-1.2.5, freetype-2.1.3, jakarta-tomcat-4.1.27 and openssl-0.9.7g all running.

My problem is this:- Apache starts and runs and all the above packages can be used. However after a while apache reaches its maxclients limit even though I KNOW that there aren't that many clients connecting. ( I'm doing some monitoring via cricket ). I have reverted to apache 1.3.27 and back regressed the php version and back. Stopped SSL, stopped java, chcked out phpinfo, all paths look correct and all info on there looks as it should.

The best way I thought to try and find out the problem was to take each package out one by one and eliminate each one that way. I have done this and am now starting to think that there may be a bad script on one of my pages but I have literally hundreds maybe even thousands of php scripts which would be impossible to trawl through each individual one.

As I said I know its a long shot but is there anyone whos had a similar problem and been able to fix it or who has any ideas. We are running quite an important service and to have to keep restarting the web server is obviously unnacceptable.

------------------------------------------
Thanx | your welcome !!

roycrom

 

[mariocq]

Hello:

What does the access log show?. If there are a lot of clients should be a lot messages displaying the requests.

Which is the output of the netstat -an command?.

If you can´t find the origin of your problem, have you checked the possibility of setting a low value for the parameter MaxKeepAliveRequests.


Regards.

 

[roycrom]

Thanks for your reply mariocq.

I think I've finally got to the bottom of the problem. What looked like an apache prob was actually more to do with my postgresql database. Don't know how much you know about postgres but it had not been vacuumed for months which led to response times to the apache processes being extremely slow so more and more child processes were being started which were all having the same problem. I have vacuumed my database now and (fingers crossed) everything is looking a lot more stable now.

Once again, thanks for your reply, it was appreciated.

------------------------------------------
Somethings come from nothing, nothing seems to come from somethings - SFA - Guerilla

roycrom