Results 1 to 3 of 3
|
Enjoy an ad free experience by logging in. Not a member yet? Register.
|
|
-
10-23-2006 #1
- Join Date
- Apr 2006
- Posts
- 10
serious load problem [urgent please]
No log handling enabled - turning on stderr logging
read_config_store open failure on /var/net-snmp/snmpapp.conf
read_config_store open failure on /var/net-snmp/snmpapp.conf
read_config_store open failure on /var/net-snmp/snmpapp.conf
Failed loading /usr/local/ioncube/ioncube_loader_lin_4.3.so: /usr/local/ioncube/ioncube_loader_lin_4.3.so: undefined symbol: zend_hash_add_or_update
I get that when tailing apache logs.
Second...
How do I trace which user/domain is causing a server load spike constantly? I can't tell if its exim or php/httpd.
I tried these, got nowhere
ls -al /tmp|grep -v sess
I have phpsuexec enabled
-
10-24-2006 #2
- Join Date
- Apr 2006
- Posts
- 10
can someone please help me out?
-
10-27-2006 #3
- Join Date
- Oct 2006
- Posts
- 4
I wish I knew as well but here are some simple thoughts I have.
How many websites are on your shared environment? I have 140 websites on one machine and everytime i have a load problem, I have to reboot the machine to fix the problem. All I can find so far is that PERL seems to run rampant at times. One of the things that I am working on to fix the problem is that we are currently setting up 5 other machines to split off some of these websites onto other servers so that we can isolate the (website or websites with the problem) but with too many sites on a server, the problem seems to be quite difficult to find or fix.
That being said, a decent linux server such as a P4 with 1gb of ram should be able to support up to 500* websites. The problem is finding the culprit responsible for hoging resources. If someone has a way to find and isolate load problems, as well as correct them - please let us know
Another question I would ask is: do you have a control panel such as cpanel or plesk? if so, you may be able to go into individual website logs or individual bandwidth reports and see if anything looks goofy from there. Otherwise, your going to have to review the logs from the command line which is a little more complicated.
This may be a stupid idea, but if you have a way to review your bandwidth, you should determine if the "high load" is incomming or outgoing traffic. Your machine could be under some sort of attack or it could have been compromised and is being used as a zombie to attack other machines. Looking at your bandwidth report may give you some additional insight about the problem.
*depending on website requirements and traffic obviously