WatchDog on CentOS...

[user505]

Hi All,

I am new in linux OS.

In my setup we have CentOS 5.3 installed...and there are 2 watchdog process running on ...and I dont know what is that?

I do not find any help what is watch dog? and their purpose..

Can any one suggest me or provide some good link so that I can identify yes, it is firewall or its at hw level or sw level.

Thanks...

[vickey_20]

whatch dog are logs

[Rubberman]

They monitor the CPU cores. My guess is that you have a 2 core CPU, correct? I have 2 quad-core xeons so I have 8 watchdogs running. When one fails, it can be "disabled" so the OS will not try to schedule jobs on it. It will let you know if that happens so you can do a clean shutdown of the system and repair it.

[user505]

vickey_20: If this is log then please let me know how to view this log?
Rubberman: yes I have 2 core cpu...but there are 2 watchdog process runnig on.

ps -ef | grep watchdog
root 4 1 0 Jun12 ? 00:00:00 [watchdog/0]
root 7 1 0 Jun12 ? 00:00:00 [watchdog/1]

but did not understood rest of your content that is for job scheduling, so can you please give me more idea on this?

Thanks...

[Rubberman]

On a single-core, single-cpu system, the operating system (Linux/Unix) will determine what the next job is that should run, and when the currently running processes timeslice is finished, it saves the running process registers, program counter (pc), etc. and loads the saved ones for the next process, continuing execution at the pc for the newly active process. In a multi-core and/or multi-cpu (smp) system it does the same thing, but apportions the processes to available cores. This makes for a quite a bit more complicated kernel. One of the things that the kernel needs to do in an smp system is to monitor the health of the cores and/or cpu's. If one goes wonky (that's a technical term for "not working right"), then it can be taken out of the pool of available processors on which it can run system or user processes. It adds quite a bit of reliability (in theory at least) to the system since it can continue to run even if one or possibly more CPU's die. Each core will have a watchdog process that periodically monitors the health of its charge, and informs the operating system to take it out of service if aformentioned wonkyness is detected. That's about as clear as I know how to make this subject here. So, you have 2 cores, hence you have 2 watchdogs. I have 8 cores, hence I have 8 watchdogs - one for each core.

[user505]

thanks for given understanding.

Can you tell me one thing why it display "0" on 4th column of ps -ef command?

And also from where I can change their value?

Thanks...

[Rubberman]

I think the C column is the CPU utilization when you ran PS. You cannot change the values of any column in PS. It only provides process information that it gets from the kernel. What do you want to do?