It's fooked, that's what. 😒
Apache starts running hundreds of processes and eventually ramps the load way up (like 80+) until it hangs and dies and is restarted by the service monitor. I have no idea why this is happening.

I figured that much... No gory details for us though?

Well, I think I added the gory details after you posted that. 😀
Anyway, I put a really low limit on the CPU seconds allowed for httpd processes. We'll see what catches.

What causes Apache to spawn additional processes? Is it a DoS attack or something?

Server usage causes Apache to spawn additional processes.. I'm not sure how it's supposed to work, but I thought that it was supposed to be that the number of Apache processes is respective to MinSpareServers and MaxSpareServers, not MaxClients. (MaxClients is 160, in which case having that many processes would make sense -- but who in their right mind would want 160 Apache processes?)

isn't maxclients the amount clients handled by an apache process before it's forcefully killed ? (useful against memory leaks)

That's MaxRequestsPerChild, which is currently set to 500.
I really wish I could figure this out once and for all, it's getting quite ridiculous. The main apache process always shoots up and causes everything to die, it's never a child that does it as far as I can tell. But, I have no way of debugging it, because the server ceases to respond to basically any command and trying to get a backtrace using gdb would take next to forever with a 30+ load. I bet if I went up to Apache2 and changed things to use suphp instead of the crappy and unmaintained phpsuexec patch things would work better, but who knows. 😒

Could it be something in the search function? I noticed it hung up on me earlier today right after I had did a search. There was a question in the DOS section about Ultima 8, and I knew the answer but also knew it had been posted before. I did a search on Ultima 8 just to check my memory and shortly afterwards got the SQL error message. It could be coincidence but wanted to mention it anyway in case that was something that could be a possible cause.

Search in phpBB is naturally weird, I don't think that was the problem. I did run REPAIR on all the tables just in case.

Still having problems posting, searching for new posts, and reloading threads as of a few hours ago (but working at the moment obviously).

Snover: So why not bite the bullet and use newer versions if they're more stable and/or less buggy? *nudge*

Because this server has a retarded front-end on it that likes to break shit like that. 😒 I'm ready to reload the stupid thing with Debian and let it loose. The only problem is my lack of experience configuring mail servers. Apache is a breeze.

can't you copy the current settings for the mail server ?

No, I would be using qmail instead of exim.

So... The whole zetafleet.com site was pretty much down all day today, saying that there were too many connections open to the SQL server...

Yup

MySQL hung itself. Don't really know why. Couldn't really tell, 'koz it was hung. It was very weird.

I wonder if there's some exploit running around for MySQL 4.0.

Dunno, but it sucks. Any plans, Fearless Leader?

How are the downloads handled? I noticed that the links go through PHP. I've also noticed thousands of downloads on some of the files. Does this add any extra load? Perhaps adding links to an FTP server or http would be better?