RE: Reentrancy from pass scripts

Thanks, Dave.

OK, so the reentrancy thing sounds. a red herring for the question I asked. *Although*-- if you actually know what --enable-reentrant would do in my environment, I'd appreciate hearing.  I'm thinking this may solve other problems I have with my cacti polling on machines with certain subsystems disabled and yet queried, resulting in proxy timeouts.

The C API thing is powerful for sure, the problem there is I'd need to do a wholesale redeployment of the agent package (we try to maintain one, network-wide, for sanity-control reasons) on behalf of the specific application server this need addresses.  That might be a non-starter operationally.  It definitely meets the dictionary definition of a pain in the behind.

I'll stick with the fact I can do a fortunate snmpget from the alternate proxy port in my pass scripts for now.  I'll look into the disman expression stuff--I'm not sure it's quite powerful enough to do all that I need even in the RFC form I remember (e.g., I need string concatenation), but if the innards seem incomplete in a way I can conceivably remedy them with some recoding within the agent, maybe I'll pitch doing that work for you myself.

Thanks again!
Wayne

-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Dave Shield
Sent: Friday, May 15, 2009 3:38 AM
To: Tackabury, Wayne
Cc: [email protected]
Subject: Re: Reentrancy from pass scripts

2009/5/14 Tackabury, Wayne <[email protected]>:
> I'm writing a pass script which will effectively synthesize a gauge and
> string object, to effectively detect a condition worth trapping on via a
> snmpd.conf 'monitor' directive.

Hmmm...   that sounds suspiciously close to the functionality of the
DisMan Expression MIB.    There's the basis of an implementation of
this in the Net-SNMP agent, but it's been a while since I looked at it,
and I can't remember how fully funtional it is.
   But it might be something worth looking at.


>                        So the point is,  my pass scripts need to
> GET other object in the management scope to make dynamic and conditional
> decisions on to drive the value of the OID's that the pass scripts
> themselves expose.
>
> The problem is-if I actually do a snmpget from the pass script, it times
> out.  ....   I'm trolling for a general understanding of how
> to fix this the "next" time or when part of what I need is from, say, the
> ucd-snmp MIB or something whose instrumentation is built into the agent.

I suspect that you're probably pushing the boundaries of what is sensible
to address with the "pass" mechanism.   If you need to perform "internal"
agent queries, then you might be better of handling this using a C-based
MIB module within the agent itself.

   That way you can take advantage of the "internal query" routines
provided by the 'utilities/iquery' module.   This is exactly what the
DisMan Event MIB does - see the code under 'disman/event' for
examples of how to do this.



> -- is this some kind of Bad Thing for pass script usage anyways?

That would be my call, yes.

>  If so, can you folks suggest an alternate strategy?

See above.

Dave

------------------------------------------------------------------------------
Crystal Reports - New Free Runtime and 30 Day Trial
Check out the new simplified licensing option that enables 
unlimited royalty-free distribution of the report engine 
for externally facing server and web deployment. 
http://p.sf.net/sfu/businessobjects
_______________________________________________
Net-snmp-users mailing list
[email protected]
Please see the following page to unsubscribe or change other options:
https://lists.sourceforge.net/lists/listinfo/net-snmp-users