[fm-discuss] Re: FMRI Reference Documentation?

[Keith M Wesolowski]

On Mon, Sep 25, 2006 at 08:32:41AM -0400, Cibrario, Robert J. wrote:

> We're looking to deploy S10U2 and need to provide a measure of hardware
> monitoring and alerting. The ability to route Fault Manager (alerts) out
> through use of System Manager Agent is most welcome. However, the
> TechOps folks who tend after our NetCool management infrastructure need
> to know how to handle such events. They need to know how to interpret
> the resulting traps so they may route the information appropriately. A
> reference of (MSG-ID) codes to be expected along with a severity
> assignment is needed. With reference in hand I can probably assign
> severity if needed.
> 
> What was the intention behind providing such a reporting capability in
> S10U2 without such a (stable) reference guide?

The message codes made available via SNMP are the same codes logged to
syslog or displayed by fmdump(1M); they're as stable, or unstable, as
they are any other way you get them, and I think Gavin's already
spoken to this more general problem.

We could, without too much difficulty I believe, add some more fields
to the MIB, such as severity.  This data we would need is available in
the dictionary files, and could be provided to the MIB plugin in the
same way as we currently provide the URL, using libfmd_adm.  It is, of
course, already possible for a human to retrieve all of this
information from the URL provided by the trap itself; am I correct in
assuming that you'd like this from the MIB in the event that your
network is not functioning?

-- 
Keith M Wesolowski		"Sir, we're surrounded!" 
FishWorks			"Excellent; we can attack in any direction!"