Kernel Error Flooding My Logs

Carlwill · 2010-01-29 00:11:22

I have a server in my home I am using for personal use and I was checking logs on the machine and saw this in /var/log/error.log

Jan 24 16:58:38 ion kernel: iTCO_wdt: failed to reset NO_REBOOT flag, reboot disabled by hardware
Jan 25 20:56:59 ion kernel: iTCO_wdt: failed to reset NO_REBOOT flag, reboot disabled by hardware
Jan 25 21:05:18 ion kernel: iTCO_wdt: failed to reset NO_REBOOT flag, reboot disabled by hardware
Jan 26 23:48:09 ion kernel: iTCO_wdt: failed to reset NO_REBOOT flag, reboot disabled by hardware
Jan 28 18:39:00 ion kernel: iTCO_wdt: failed to reset NO_REBOOT flag, reboot disabled by hardware

I did a search on the forums and Google & couldn't find any answers. Does this make sense to anyone?

[carlos@ion ~]$ uname -a
Linux ion 2.6.32-ARCH #1 SMP PREEMPT Sat Jan 23 11:41:48 CET 2010 x86_64 Intel(R) Core(TM)2 CPU 6300 @ 1.86GHz GenuineIntel GNU/Linux

res · 2010-01-29 00:47:09

itco modules are for watchdog, a utility that shutdowns the system if a file/fifo doesn't get periodically written to.

Do: lsmod | grep -i itco ...and blacklist the modules in rc.conf if you don't use watchdog.

No harm in not loading them; Debian does this by default.

Aedit · 2010-01-29 00:51:30

It's from the kernel driver for the Intel total cost of ownership (TCO) watchdog timer in the I/O controller hub on your motherboard. That hardware would try to detect and resolve serious software or hardware problems (lockups). It's saying that it would not be able reboot in event of an error. Don't know why you are seeing the message though.

Carlwill · 2010-01-29 01:06:03

res wrote:

itco modules are for watchdog, a utility that shutdowns the system if a file/fifo doesn't get periodically written to.
Do: lsmod | grep -i itco ...and blacklist the modules in rc.conf if you don't use watchdog.
No harm in not loading them; Debian does this by default.

[carlos@ion ~]$ lsmod | grep -i itco
iTCO_wdt                8069  0 
iTCO_vendor_support     1841  1 iTCO_wdt

So to 'blacklist' the module in /etc/rc.conf I assume I would 'uncomment' the section that says "#MOD_BLACKLIST=() #deprecated" but what is the specific name of the module I would need to add into the rc.conf file? Is it "itco".

Aedit · 2010-01-29 01:11:38

In rc.conf it says

# NOTE: Use of 'MOD_BLACKLIST' is deprecated. Please use ! in the MODULES array.

so do as it says and put

MODULES=(!iTCO_wdt !iTCO_vendor_support)

res · 2010-01-29 01:11:44

MODULES=( ... !iTCO-wdt !iTCO-vendor-support)

e: what he said ;)

Last edited by res (2010-01-29 01:13:14)

Carlwill · 2010-02-06 00:03:16

I have had this in my /etc/rc.conf file for about 1 month:

DAEMONS=(syslog-ng network netfs crond sshd ntpd !iTCO_wdt !iTCO_vendor_support)

I just happened to check my logs again this Friday while we perform updates on the Dell servers and I still see:

[root@ion log]# tail errors.log
Feb  5 18:38:32 ion kernel: iTCO_wdt: failed to reset NO_REBOOT flag, reboot disabled by hardware
[root@ion log]# lsmod | grep -i itco
iTCO_wdt                9093  0 
iTCO_vendor_support     1841  1 iTCO_wdt

Why is this module still being loaded if my rc.conf file is telling it not to. Did I do something wrong?

jowilkin · 2010-02-06 09:40:36

You need to blacklist them in the MODULES array, not the DAEMONS array.

Arch Linux

#1 2010-01-29 00:11:22

Kernel Error Flooding My Logs

#2 2010-01-29 00:47:09

Re: Kernel Error Flooding My Logs

#3 2010-01-29 00:51:30

Re: Kernel Error Flooding My Logs

#4 2010-01-29 01:06:03

Re: Kernel Error Flooding My Logs

#5 2010-01-29 01:11:38

Re: Kernel Error Flooding My Logs

#6 2010-01-29 01:11:44

Re: Kernel Error Flooding My Logs

#7 2010-02-06 00:03:16

Re: Kernel Error Flooding My Logs

#8 2010-02-06 09:40:36

Re: Kernel Error Flooding My Logs

Board footer