Geeklog Forums

Has anyone had a problem with geeklog causing run-away Apache HTTPD processes that end up taking the processor to 100%? I'm watching my server right now and the CPU usage is sitting near 98% and apache has been active for 388 minutes. Is there a problem between geeklog and certain versions of Apache? This only happens after I spend a fair amount of time working on the site, and pretty soon I'll be getting memory allocation errors again as it slowly drinks up all of the available RAM.

Hoo-ee, it looks like I'd better wait for some more information on this topic before I try installing GeekLog on my web host's server. They get mighty touchy about folks who hog the server resources, for any reason at all. Hope someone will post soon with some clarification on this. Once the "all clear" is given, I would really like to try out this very promising application. Cheers!

I can assure you that this is certainly not the normal behaviour of Geeklog. I have installed Geeklog several times on different platforms and never experienced problems like this. Besides, my web hoster would certainly have kicked me by now if Geeklog misbehaved like that ...

As usual, the best idea is to try it out yourself. But as far as I can see, this is an isolated case.

bye, Dirk

First, let me say I agree with Dirk, this is not something that we would expect, or anyone else should expect from geeklog/apache. Having said that, before we can do you any good please let us know a little more about the environment that this is set up in. What OS you're running, what version of apache, what version of php, if php is a module or cgi, etc. It'd be even more helpful to us, if you have access to the machine (and I'm hoping since you have top or ps you do), if you can configure apache to use server-info and tell us what processes are taking the most time for apache. It'd also help for you to explain what you mean by "working on the site", what exactly you're using and doing. If memory starts going up that normally means, in my world, that there's a memory leak. If you're always viewing your pages with an older version of mozilla or netscape for instance on linux, that could be the source of your memory leak. Even depending on what you're using to monitor your system (GTop, ROX system monitor) could be leaking (though I'm not accusing them). If you are using a system monitor you might cut in the important sections to a comment so we can see more details. I know I haven't solved anything yet, but hopefully you'll give us the tools to help you. The last thing we want from geeklog is to find out that when certain planets align (and i'm in vegas on vacation) we just start chewing resources.

Heya, First off, I didn't mean to give the impression that GL is the definite source of the problem, I was just wondering if anyone else was experiencing this. I have installed Geeklog on many machines in the past and this is the first time I've ever seen this happen and it just happens to happen when a geeklog site is accessed a lot. The machine the site is running on is an RS6000 running SuSE Linux, 2.2.14 kernel. 375MHz PPC processor, 512MB RAM. Apache version 1.3.20, PHP 4.0.6 (as a module). When I say working on the site, I mean making modifications and adding stories within the geeklog system via a web browser. I would include a clip from top of the processes going nuts, but the system admin restarted apache this morning because there was a process running for around 1500min as well as several other HTTPD processes that had been running for several hundred each. Here's one line from a screen of top I took about a week ago: 9516 nobody 10 0 5024 5024 2508 S 0 28.9 15.5 456:34 httpd Now, add on to there about 4 other processes using varying amounts of CPU, RAM and had been up for several hundred minutes. The machine is a university computer that I don't have direct access to. No one actually uses the machine directly, it's a development box that a bunch of students use to host web pages and projects for classes. Currently however there are only two sites running on it (both geeklog), which led me to think it may be a compatability problem. The system admins were toying with the idea of reinstalling apache and php to see if it was a problem with how they configured it in the first place. Any more info needed? -Mike ----- Linux - Because rebooting is for adding new hardware.

Well, first I'd upgrade PHP if for no other reason than for getting some recent security patches now that we all know you have an exploitable version of PHP ;-) Second, can you confirm that when GL isn't running that this does not happen? The closest thing I ever had to a resource problem with Geeklog was on Windows. Windows doesn't use the mysql_pconnect properly so I changed that to just mysql_connect and it worked fine. Outside of that, my guess is your problem is outside the realms of Geeklog. dreamscape may be able to offer futher advice. ----- The reason people blame things on previous generations is that there's only one other choice.

---

The reason people blame things on previous generations is that there's only one other choice.