strange memory problem

Hello,

.. i got an apache / RedHat server with about 320 MB ram and MySQL and php running.

Last weekend i got memory problems and the server began to use the swap memory. Unfortunately this brought the server down and due to that memory "leakage" i made an upgrade from about 190 MB rams to 320 MB, like the support of my hosting company suggested.

My server isn´t a real heavy loaded one but i got about 100.000 hits a day.
The real problem is that the 320 MB are meanwhile not even enough ! During the last week the memory consumption increased every day and now the server is going to use the swap memory again....

If i stop the apache server the used memory stays the same, if i stop the mysql server nothing changes too...

...striking is that about 250 MB are used for buffers

Does anybody has experience with that kind of problems ?!?

Any suggestions would be useful !

Thanks


Wolfgang

Set the following in httpd.conf, it should be more than enough to handle 100,000 hits.

StartServers 150
MinSpareServers 50
MaxSpareServers 100
MaxClients 512
MaxRequestsPerChild 30000
rlimitcpu 90 90
rlimitmem 40000000 40000000

Go to http://httpd.apache.org/docs/mod/directives.html to check out each of them. If at all possible, don't use Redhat (enables so many daemons by default), try any *BSDs.

thanks for the advice, but it doesn´t help...

If the memory leckage is a apache problem, why doesn´t it help to turn on/off the apache server ?

>> why doesn´t it help to turn on/off the apache server ?

Why should Apache turn on/off itself? Not to mention Apache, any daemon that reaches its max resource level would crash. You should recompile your kernel and adjust/optimize your system. In *BSDs, you can specify how much resource you wish to allocate for TCP. If you can't do it in Linux, too bad and live with it.