UBB.Developers Forums Community Development Programming MySQL and Server Configuration Runaway Apache processes

Late last night my apache processes went haywire and started multiplying like crazy, reaching the maxclient setting (of 150) and always running making the load go to insane levels like 250! The log shows that 10-20 processes dies (seg faults) after the maxclient is reached. Have searched Google but can't find anything that applies really.

I need help badly with this...

What I'm using:
Red Hat 7.1 (with almost all upgraded rpm:s)
Linux 2.4.18
Apache 1.3.27 (no ssl)
PHP 4.2.3
MySQL 4.0.6

I wish I had some input to offer on this. My usual solution would be to reformat and start over... LoL

That's the windows way to handle things.

I've found some info which makes me believe that it is a hardware problem, most likely the memory. I will try that as soon as possible, but the server is about in another city so I'll have to wait until someone comes in at work tomorrow morning to try and fix it. It's a pain to not have direct access to the server when it breaks down. =[

LoL I've done the "Windows Way" of fixing on my Mandrake install a couple of times. Heck... I kinda like the nice new car smell a fresh OS has. LOL

I'm planning on installing Debian some day, but first I'll have to run it at home for a while to make me comfortable using it.

Have been planning to do that for quite a while now, should get started on it soon I think. =]

I have had this problem on ours a few times and it was always MYSQL. Usually because the server had high traffic causing mysql to "screw up" (So technical I know..)

To do an immediate fix I would just telnet in as root and do a killall on mysql and then restart mysql.

A Long term solution was to upgrade the RAM to 256.

Hope this helps,

-Jason

Sadly no. The mysql-processes are fine, restarting them makes no difference, and we've got fairly much memory (896 MB). The problem is apache processes starting to crash (segmentation fault, trying to read past its own memory area) for no reason and generally start acting weird.

According to some sources it is almost certain that this means there is a hardware problem, as Apache doesn't have any memory access problems. Probably faulty, or inproperly installed, memory is causing this. Unfortunately I still haven't been able to get to the server to actually try it out yet.

Since we just found out what was wrong I thought I'd inform you.

It was one of the sites I'm hosting that had written a script that went into a never ending loop (because it coudln't handle incorrect input) which hogged all apache processes.

I don't really know why PHP didn't kill them properly, since max_execution_time is set at 180.

It wasn't easy to find that bug, especially since the main programmer of that site told me several times that they hadn't done any changes at all lately (which one of his minions actually had...).

Now I'm happy again. =]

Thanks for posting the answer.

In my earlier days of programming, when I used mainframes or embedded systems, I knew exactly how to diagnose such problems. With internet programming, the tools I'm accustomed to don't seem to be available.

Nope it's a real mess tracking down problems like this. I don't I would have made it myself, so I'm really happy for the help I got.