Hello,

After latest v3 update, cc-control process under ccuser, is eating all CPU power on the server.

Every time I restart Centova everything is back to normal, but after couple of hours cc-control is hungry again and start eating all CPU.

When looking at what cc-control is doing, this is what strace return (repeating):

Code: [Select]

select(0, NULL, NULL, NULL, {...})      = -1 EFAULT (Bad address)

Currently I'm having prm watching cc-control and it restart Centova on high CPU usage and execution time.

Regards

Hmm, interesting.  Thanks for the strace output -- very helpful!

select() under Linux should not return EFAULT.  Googling for this yields a few relevant posts (most notably from the Apache development mailing list, as apparently this affected Apache too) in which it was speculated that this was some kind of long-standing kernel bug.

I don't know what the cause is, but I've added some code to protect against this and sent an updated version to the download server.  Let me know if that resolves it for you.

Every time I restart Centova everything is back to normal, but after couple of hours cc-control is hungry again and start eating all CPU.

Similar thing was in my case but it was after day or two rather than hours. After updating kernel to 2.6.32-279.22.1.el6 the problem seems to be disappeared.

Great, I have update the to latest build and will monitor CPU usage for couple of days.

It's very possible that kernel is causing this error. It's custom build from OVH for their CentOS 6 install image (3.2.13-grsec-xxxx-grs-ipv6-64).

If problem persist, I will try to manually update the kernel.