Perhaps you guys could shed some light on this. Yesterday, some users started reporting their sites were very slow to respond. All tracerts and pings showed normal. Running top in the server showed no unusual processes and everything looked normal (did notice probably fewer http processes than normal).

So, if the connection to and from the user is fine going off of tracert and the server itself isn't showing a high load, where would be the next place to look? We have multiple servers that and this server is a slave. All other servers appear fine and all sites load like normal. Just this one servers sites are very slow to load. Any suggestions?

Perhaps he's running a badly written script?

Did the user happen to be retrieving his files via ftp at the same time?

This is serverwide and not specific to a particular user. When running 'top' everything seems fine and I would think that poorly written script would show something was going on. Server load, etc seem fine.

Well. One of my users file a ticket (which we can't exactly duplicate) that when he is backing up his files daily, via ftp, the site starts to crawl. This does show on Top or anything because it's probably i/o issue.

But that said, we didn't managed to replicate the problem. So it belongs to only the "Suspected Issues" category currently. But perhaps that may be your cause.

That very well could be. Is it possible to tell if somebody has something going on via FTP. We have a lot of clients that use datafeeds for affilate programs so they are moving a lot of data. Usually it isn't a problem but this time it makes me wonder. I'm not even sure where to look to see who is doing what, however, what you describe sounds like it very well could be our problem.

You can check the processes. It will show. But you can't exactly tell what they are doing although xfer_log will show if they are transferring a huge amount of files.

I don't see anything out of the ordinary.... The only think I don't recognize is one single line. Could be normal, I just don't recall ever seeing it.
----------
7 root 9 0 0 0 0 SW 0.0 0.0 0:00 scsi_eh_0
----------

Here is the rest of my top.

8:06am up 2:51, 1 user, load average: 0.12, 0.08, 0.05
227 processes: 226 sleeping, 1 running, 0 zombie, 0 stopped
CPU states: 0.9% user, 0.5% system, 0.0% nice, 98.4% idle
Mem: 515676K av, 466916K used, 48760K free, 0K shrd, 90040K buff
Swap: 530104K av, 716K used, 529388K free 194936K cached

PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND
12675 root 17 0 1144 1144 824 R 0.7 0.2 0:02 top
862 nobody 11 0 1304 1304 1072 S 0.1 0.2 0:00 proftpd
12735 nobody 10 0 9808 9808 9392 S 0.1 1.9 0:00 httpd
1 root 0 0 476 476 416 S 0.0 0.0 0:06 init
2 root 9 0 0 0 0 SW 0.0 0.0 0:00 keventd
3 root 19 19 0 0 0 SWN 0.0 0.0 0:00 ksoftirqd_CPU0
4 root 9 0 0 0 0 SW 0.0 0.0 0:00 kswapd
5 root 9 0 0 0 0 SW 0.0 0.0 0:00 bdflush
6 root 9 0 0 0 0 SW 0.0 0.0 0:00 kupdated
7 root 9 0 0 0 0 SW 0.0 0.0 0:00 scsi_eh_0
8 root 9 0 0 0 0 SW 0.0 0.0 0:00 kjournald
141 root 9 0 0 0 0 SW 0.0 0.0 0:00 kjournald
142 root 9 0 0 0 0 SW 0.0 0.0 0:00 kjournald
143 root 9 0 0 0 0 SW 0.0 0.0 0:00 kjournald
144 root 9 0 0 0 0 SW 0.0 0.0 0:00 kjournald
145 root 9 0 0 0 0 SW 0.0 0.0 0:00 kjournald
712 root 9 0 548 548 460 S 0.0 0.1 0:00 syslogd
717 root 9 0 448 448 388 S 0.0 0.0 0:00 klogd
879 named 9 0 4648 4648 1992 S 0.0 0.9 0:00 named
881 named 9 0 4648 4648 1992 S 0.0 0.9 0:00 named
882 named 9 0 4648 4648 1992 S 0.0 0.9 0:08 named
883 named 9 0 4648 4648 1992 S 0.0 0.9 0:00 named
884 named 9 0 4648 4648 1992 S 0.0 0.9 0:01 named
904 root 9 0 1216 1216 1100 S 0.0 0.2 0:00 sshd
937 root 9 0 868 868 728 S 0.0 0.1 0:00 xinetd
974 root 8 0 2608 2608 1404 S 0.0 0.5 0:00 chkservd
1007 root 18 19 4892 4892 1596 S N 0.0 0.9 0:03 cpanellogd

Now it looks fine... You have to do it when the system is slow :)

Another possibility is that your memory is maxed out, and that the system is doing a lot of swapping. That can also slow down your system and won't show on CPU loads, and is more of an I/O issue.

The thing is what top is taken when it was slow. Most sites are just crawling. One thing to note, notice the lack of http processes. There are few that pop up, but not many. Apache seems to be running fine (at least it shows that in WHM and restarts fine). Rebooted the server this morning and still same issue. That top was taken almost 3 hours after re-boot.

How many processes and child you have in your httpd.conf?

I will take a look at that..


One thing that is standing out to me right now is the length of each process. Watching top run, nothing lasts more than 0:09 if even that long. Looking at another server, many processes will run for various lengths of time.

Try

Maxclients = 512
Maxrequestperchild = 100
MinSpareservers=5
MaxSpareservers=10
Startservers=5

Can play around with those.