High Load Average

Hello,

I'm using ISPConfig3.0.4.5 with Debian. The server has 4Gb RAM with Xeon QuadCore.

I always have a high load average over 2,00 with peaks of 8,00 -10,00 or 14,00

The CPU works more or less 5%
The memory status is:
Mem: 4138076k total, 3787252k used, 350824k free, 351980k buffers
Swap: 6067192k total, 3296080k used, 2771112k free, 1376088k cached

The most time the disk is busy for 70-100%.

the service that uses more disk is imapd, then kjournald, du, and cleanup. Sometimes mysql uses disk recourse but is punctual.

The server has and average of 10 msg/min sent and 15 msg/min received

To decrease the amavis server use i've this configuration on ispconfig:

Virus lover-> Yes
SPAM lover ->No
Banned files lover->No
Bad header lover->No
Bypass virus checks->Yes
Bypass banned checks->Yes
Bypass header checks->No

Could you give me some idea? what can i monitor to understand what is happening?

Thankyou in advance

 

I have high load overage too.
I'll stay here to see what people can tell us.

 

Do you use fast disks (7,200 RPM or more) or an SSD? Do you use RAID0, 1, or 10?

 

I'm using raid1 with two Sata disks partitioned with ext3

Thank you.

 

That should be fast. Do you have Munin installed on the server? It can help you find out where the bottleneck is.

 

Hello Falko,

Thanks for your indications.
Finally, after severals days monitoring, we have decided to migrate the mail service to another server.

Actually we have one server with a single server installation.

We want to install ispconfig with mail service to a new server and then move the mail domains and all the email accounts from the present server to this new one.
Is there any tutorial to do this migration?
if this possible if we have, in this moment, a single server installation on our server? exists an automatic way to move services between servers?

Thank you

 

Hello Falko,

Something strange occurs. If we stop the /usr/local/ispconfig/server/server.sh execution in crontab -e, the server works fine in reference to load average. When we put it again every minute the server load average grow up with peaks of 25 and 35.

Now we have solved , more or less, the problem, modifying the crontab executing the server.php script, every 30 minutes, In this way the load average only grow up every 30 min.

Some suggestion?Hello Falko,

Something stange. If we stop the /usr/local/ispconfig/server/server.sh execution in crontab -e, the server works fine in reference to load average. Wen we put it again every minute the server load average grow up with peaks of 25 and 35.

Now we solve more or less the problem modifying the crontab xecuting the server.php script every 30 minutes, In this way the load average only grow up everey 30 min.

Any suggestion?

 

Hello Till, thanks for you reply.

- Do you use courier or dovecot on the server? -> Courier
- Is the directory /var/vmail were the emails are stored a network drive, e.g. nfs or glusterfs or similar? No. All HD is one partition in ext3
- Have you run the server.sh script in debug mode as described in the faq to see were the bottleneck of your installation is? Yes. The result:


- DEBUG - Set Lock: /usr/local/ispconfig/server/temp/.ispconfig_lock
- DEBUG - No Updated records found, starting only the core.
- DEBUG - Remove Lock: /usr/local/ispconfig/server/temp/.ispconfig_lock
finished.