Hi,
Well, this is not a development project, instead it is a problem evaluation task. We have a Server which works for Domains e also for Mail (pop and smtp) of these Domains. Something around 700 domains and 2600 mailboxes.
The Domains are not CPU or Band consumers, they are all from small companies, static sites, very few applications. The Mailboxes, well, they are mailboxes and you can imagine what this means.
Since from a couple of weeks ago this Server is presenting very very high Cpu Load Average Times. It normally worked with values from 0.x to 15.x, what was something really normal for us and the Server supported it nicely, without bottlenecks. But now (it is intermitent, not always), mainly during the work hours of the day, it has gone to values as high as 200.x, 270.x CLA times. And when this happens, of course, we have real troubles to read / send emails, to FTP, to access sites, everything is refused.
So, I would like someone to find out and helps to solve this problem. I figure it in at least two steps:
1) Examining the following files, with data selected from those peak periods:
/var/log/maillog
/var/log/messages
/var/log/httpd/error-log
/var/log/dmesg
Maybe this could be enough.
2) But if it is not enough we could plan and arrange so that the Server can be examined exactly during one of these peek periods. For this we could work together, maybe typing at MSN or ICQ.
## Deliverables
1) If there is the need to inspect directly the Server when the problem is happening, we must agree a way of get in touch immediately and work together, while we type through MSN or ICQ.
2) At the end of the examination, if the causes of the problem are identified, we want to have a small report with the description of the problem and the recommendations to solve it.
3) When we accept the bid we shall attach the LOG files mentioned in the description.
## Platform
Linux Centos Server
Pentium-IV
3.x Mhz
1.5MB RAM
HT Processor
SCSI HD 72 GB
using:
Apache Server
Sendmail
Qpopper
Spam Assassin
MySql