topic Re: Smartmontools in Operating System - Linux

Smartmontools

wobbe — Tue, 02 Feb 2010 22:45:30 GMT

So, how good/usefull is smartmontools?

I installed smartmontools on a DL380G2 running Debian Lenny. I have 6 harddisks so I configured /etc/smartd.conf for 6 disks.
I've setup email and run smartctl for all disks to check their status. All disks are fine.

Now do run some tests and I pull one of the disks.
Running a status report with smartctl I find out that it's no longer showing me any information for disk 5. But I did not pull disk 5, I took out disk 3! So any smart warnings for disk 5 would have gotten me to replace disk 5 where infect I should have replaced disk 3. (I'm ignoring any status lights)

Hoping a restart of the smartd would trigger an allert I now stop the service. But now I notice that I can't start the service anymore.
As long as there is failed/missing disk the smartd won't start. It will also not start after a reboot.

So in case of an abrupt failure I don't get an email warning and after a restart I also lose smart monitoring for the rest of the disks.

So is smartmontool any good? Or should I start looking at other tools? And what other tools will work (stable) on Debian Lenny with the cciss controller?

perhaps try one of these:
http://cciss.sourceforge.net/#cciss_utils

Re: Smartmontools

Goran Koruga — Wed, 03 Feb 2010 06:06:05 GMT

Hello.

How did you configure the disks in smartd.conf?

Goran

Re: Smartmontools

wobbe — Wed, 03 Feb 2010 08:49:22 GMT

I disabled the DEVICESCAN and have six time the line below.

/dev/cciss/c0d0 -d cciss,X -a -s (L/../../7/02|S/../.././02) -m emailaddres@compagny.nl

(X = disk number 0-5)

Re: Smartmontools

Goran Koruga — Wed, 03 Feb 2010 09:36:51 GMT

Right, that would explain why it doesn't start.

Perhaps using '-d removable' will help.

Goran

Re: Smartmontools

wobbe — Wed, 03 Feb 2010 09:57:00 GMT

Hi Goran,

Are you suggesting replacing cciss with removable, perhaps adding removable between -d and cciss?
I've tried both but then the service wont start.

WB

Re: Smartmontools

Goran Koruga — Wed, 03 Feb 2010 10:27:16 GMT

Did you try it this way:

-d removable -d cciss ...

Goran

Re: Smartmontools

wobbe — Wed, 03 Feb 2010 10:55:57 GMT

That's better. I can now restart the smart daemon when a drive has been pulled.
Is there perhaps an option to get an email when a drive is missing/defective?
And how do I get those smartd.conf disk numbers to correlate with the numbers on the drive cage?

Re: Smartmontools

wobbe — Thu, 04 Feb 2010 20:55:24 GMT

It doesn't matter what hardrive you pull it's allways failing the smartctl check for number 5.
So for now I run the attached script every 15 minutes and that seems to work when a disk is missing/defect.
If anyone had a better idea or thinks there is a problem with this setup then please let me know.