CPanel Update notice -extra cpu's

[embsupafly]

I saw this in the printout of a WHM update:

==> WARNING: configured processor count does not match
==> actual processor count (4)! Running stats programs
==> on this system may cause excessive load!
==> You should set extracpus=0 in /var/cpanel/cpanel.config if this is not ok.
==> cPanel Log Daemon version 22.2
==> Shared RRDTOOL support enabled

We do have 4 cpus, and loads still get high when stats and backups run. What I don't understand is, is it trying to tell me that I should tell CPanel that I only have 1 cpu (by entering extracpus=0 ?) Why would I want to tell it I have less CPU's than I actually do? Will it affect the extra cpus on anything besides stats?

[embsupafly]

actually, I looked in my /var/cpanel/cpanel.config file and it says:

extracpus=20

Any reason why. Since I have 4 total, should I change this to 3?

Do I then have to restart cpanel?

Would this explain high cpu loads for stats and cpbackups? CPanel is taking up more cpu time because it thinks it has 20 extra cpu's?

[IRCBrasil]

same problem here.

Someone knows why?

[carock]

When the nightly backups run, they will use the value of extracpus to determine at what load average of the computer to stop the backup jobs and wait for the load to go down.

In the old days, a load average of 1 was the max for a single cpu system, and a load average of 2 for a dual cpu system. At least that was my experience.

Now days, the load average is an "average" of many things, not just cpu usage. On my Pentium 3 boxes, they run fine with a load average of 3 or 4, and are hard to use with a load average of 8 or so. With my Dual Xeon systems, they operate fine with a load average of 20 and are hard to use with a load average of 30+.

To "trick" cPanel backup into actually doing the backup jobs and not stopping every 30 seconds, you have to put an inflated number in there for extracpus. On my Xeon boxes it's like 16 or so. If I put in 2 for the actual number of cpus, the jobs would never run, that's a normal operating load for my machines.

You should find out if your backup jobs are stopping frequently by looking at thier E-mails they send when they complete. It's also a good idea to know how high the load average can go on your machine before it's unresponsive to http or smtp request.

Chuck