[hobbit] Disk monitoring feature request

[Henrik Stoerner]

On Mon, Jul 31, 2006 at 10:30:05AM -0400, shea_greg at emc.com wrote:
> So if people in the BB community are called brothers, does
> that mean the folks here are called hobbits??

Guess so :-)

> With my current version of BB, I've modified the bb-disk.sh so
> that each filesystem in error sends a page message to the server
> not just the most severe (red).  So in other words, if 2 filesystems
> are in warning (yellow) and 1 is in panic (red), I would send 3
> page messages to the BB server, one for each error.  The BB server
> would then an SNMP trap for each error in addition to any email
> notifications that are in place.  These get handled by the SMARTS.
> 
> So.... Is it possible to add this functionality to Hobbit?

Not directly, it's something you'll have to build yourself. But it
shouldn't be all that difficult.

On the Hobbit server, setup a SCRIPT for the disk alerts. This triggers 
along with the e-mail alerts you may have for disk errors. Alert scripts
are fed - among other things - the full status message of the alert,
in the $BBALPHAMSG environment variable.

So you could build a script that looked through the alert message and
triggered on any "^&red" or "^&yellow" lines - there will be one for
each of the filesystems that are critical. Then your script can send out
the SNMP traps through some SNMP utility (Net-SNMP has a snmptrap tool
to do that; there are also Perl modules for this, I'm sure).


Regards,
Henrik