Atomicorp

Those messages are harmless, they're just picking up some buggy motherboard messages. There are only two components other than the kernel that touch apache, mod_security and mod_evasive. If you've used the default kernel and your load is still going through the roof then the kernel was ruled out when you did that.

You can try disabling mod_security and mod_evasive next, just mv their conf's out of /etc/httpd/conf.d/ and restart the daemon. If you're still having problems, then you've ruled out all of ASL completely and its something else.

Then I would start looking around for the usual culprits, .htaccess files for any server related settings. I've seen mod_rewrite do it when it gets into a loop, badly written apps, etc.

I am receiving this kernel BUG now as well... I can't seem to get rid of it.

I tried apci=off and noapic, neither of them work

Mar 5 14:10:01 ch02 kernel: BUG: sleeping function called from invalid context at include/linux/pagemap.h:165
Mar 5 14:10:01 ch02 kernel: in_atomic():1, irqs_disabled():0
Mar 5 14:10:01 ch02 kernel: [<c146856c>] [<c146ae3d>] [<c1420c43>] [<c1421084>] [<c161a0bb>] [<c1455ea6>] [
<c1420c43>] [<c161b8e5>] =======================


It really causes problems with syslog, since it obscures real input... and I don't want to eliminate kernel messages.

We really need to track down the source of the problem, not mask it.

Any other suggestions?

We have a newer kernel (2.6.24.2) in the asl-2.0-testing channel currently going through QA to address this problem.

scott wrote:We have a newer kernel (2.6.24.2) in the asl-2.0-testing channel currently going through QA to address this problem.

Ahh ok, good to hear. I will deal with the log as-is for now then.