CPU Cycles from 0 to 100% on Dashboard

Hello,

We have a few FPBX 2.4.x boxes in production (some older, some newest 2.4.x) and found that after ~8 weeks of up time, the CPU flashes between 0 and 100%.

Some background:

  • Call quality is never compromised.
  • Issue is resolved by rebooting the server.
  • Does not appear to have any relation to server load (I.e. we have a dual xeon box with nothing configured load, just for testing purpose, issue persists).

What is similar:

Servers are all from Dell. We use R200, 1950 and 2950.
All servers are using CentOS 5.x
All servers have > 4GB Ram.

Can anyone help tell me why the CPU cycles like this?

Thanks in advance!

-Rick

If you run top in a terminal window, it should tell you what is eating those CPU cycles.
That’d be where I’d start looking.
Let us know what you find.

I am also have a server whose CPU will cycle from ~0 to 100% CPU.

The server is running FreePBX 2.5.1, Asterisk 1.6.0.5, and Fedora 10. The machine is a Dell 2900 server.

In my case the problem starts after the machine is rebooted. top, shows that asterisk is using up all of the cpu cycles.

I finally found out that disabling and re-enabling the Queues, stops asterisk from using up all of the CPU. Have you tried doing anything like this?