> I had a strange apache problem. > > I noticed that port 80 was not serving up web pages at the normal speed. > > Port 81 was fine (admin port). > > It took maybe 90 seconds to get a page. Lots of people complained. > It affected all virtual sites. It wasn't MySQL either. I restarted apache > and > it was ok for 1 minute > and then went back to being slow. > > I loaded the few packages (patches) that I had not > loaded yet and today its fine (so far). The CPU was more than 50% idle, so > that wasn't it. > > Could Apache not have created enough spare servers? How can I check that > condition? > > Has anyone seen this on their systems besides me? > > Herb
Herb, it's funny you should mention this, because today I had a similar problem. Mine was related to low RAM and I had to reboot. However this problem has been going on for weeks and I'm getting ready to migrate to a newer server with greater resources. Here's what I did to trouble shoot the problem: # top -bn1 >output.txt This outputs a complete list of all running processes to a file called output.txt. It includes how much ram, cpu and etc. that they use. Because it's in a nice structure it can very easily be imported into a spreadsheet such as excel. I can then very quickly resort the columns and total up fields to find who's eating up my resources. You'll see that the user column is different for the admin webserver and the normal webserver. Don't mix the two together. It may be that your problem was too many server children running. If you start doing heavy swapping you'll definitely notice some slow downs. I have modified my httpd.conf a little. I found that with 256 MB of RAM, it's very easy for a webspider or errant program to cripple my server if the default number of maxchildren is left in place. Matt Nuzum _______________________________________________ cobalt-developers mailing list [EMAIL PROTECTED] http://list.cobalt.com/mailman/listinfo/cobalt-developers
