Where ISPC says System Load 1 Minute: 9.53 System Load 5 Minutes: 5.41 System Load 15 Minutes: 3.36 What are the numbers? percents? And, here: Users Online: 4 users Those are anonymous users browsing the sites? or logged users using the ispc panel? Thanks
Have a look here: http://en.wikipedia.org/wiki/Unix_load_average That means that 4 users are logged in on the shell. That could be you using 4 PuTTY or SCP sessions at the same time. To see who is logged in run Code: w on the shell.
You can try to optimize your system, buy more RAM, etc. But as load average of 9 and above is too high. Normally it should be below 1...
Its a P4 3.0 H/T 1Gb Ram and running just 10 sites... just 1 is ussing database (9 are just html) how could it be?? i will not buy a P-D with 4 Gigas ram because 1 site !! im running a 30 times bigger site under win2003/asp in a p3/750 with 512Mb !!
In my experince linux is in most cases faster then windows when you use server applications, so there must be some other problem. 1) Which programming language does your database site use? 2) Which database do you use? 3) What is the size of the database? 4) Are you sure the high load is not caused by emails (spam scanning and antivirus check?
1) PHP 2) MySQL 3) somewhere around 5Mb 4) who knows.. how can i see it? is there something like windows's task manager?
Do you use PHP as mod_php? How many pageviews do you get approx. per day on this site? You can see the running tasks with the command "top" To quit the top program, hit "q".
Something is wrong with your configuration and you need to start troubleshooting. I have an Intel Xeon 2.8 (roughly a PIII) with 1GB RAM and 174 sites up. I am barely stressing the server and average about .6 on the server loads. You did not mention your OS. Is it CentOS 4.x???
Fedora 4.. i changed some config stuff and get a 1.3 average.. is there a way to see more information? i tried vmstat but its illegible.. i need something like windows task manager or something dinamic where i can see if the problem is postfix or apache or anything else.
First, ive found that the origin of the "Procmail cant create file" is bad rights on /var/spool/mail folder, i set them to 1777 (avoids problems) and it worked fine... till midnight, then they go back to dw--------... may be this is fine and there is an error with the postfix's user, but i can stop to look at this, so i had to program a cron job every 10 minutes (to avoid client calls) like "chmod 1777 /var/spool/mail".. it was yesterday, no problems till today. Second, i executed the "top" job... now its a relief... i dont know where ispconfig get the "9" data, but CPU usage is always under 3% as top. Maybe the "9" is cause slow hd or something, but a 3% as max usage is a lot better than %900 !! Thanks !
Ok, you mean the load average in ISPConfig, load 9.0 means 9%, the load average in ISPConfig is from the "uptime" command.
Im fanatized with the top command now... je.. i found something, there is a "awk" program using above 30% procesor time since 1 hour or more... its not fine.. i readed its a script interpreter, but its the first time i hear about it, ispconfig is using awk for something? its from httpd? do you know something about it?
There are millions of programs that use awk... ISPConfig uses it only in the procmail .quota.rc recipe (if you use mbox, not Maildir). I think it's safe to kill the awk process: Code: kil -9 <pid of awk process>
I believe the problem was awstats.. i remove it and everything appears to be normal right now... thks
