RE: snmpwalk is sending an exact request when walking a tree

> From: [email protected] 
> [mailto:[email protected]] On 
> Behalf Of mn kh
> Sent: Thursday, July 03, 2008 11:42 PM

> Following is my mib as printed by print_mib function:
> Children of mmsc(8):

	The printout appears to be counting down..?

> when i run the subagent and the snmpd and i try to walk this 
> tree this is what happens in the logfile:
>  
> var_jAgentX -> Handling an exact request...
> 10:48:56.641 ( 4858) 5 var_jAgentX -> 
> name=enterprises.jinny.generic.mmsc.msend-total
> 10:48:56.641 ( 4858) 5 var_jAgentX -> It's not a table
> 10:48:56.641 ( 4858) 4 var_jAgentX -> header_generic failed for mmsc!
> 10:48:56.641 ( 4858) 3 var_jAgentX -> Handling an inexact request...
> 10:48:56.641 ( 4858) 5 var_jAgentX -> (before) name = 
> enterprises.jinny.generic.mmsc.msend-total
> 10:48:56.641 ( 4858) 5 var_jAgentX -> (before) vp->name = 
> enterprises.jinny.generic.mmsc.msend-total
> 10:48:56.641 ( 4858) 5 var_jAgentX -> (after) vp->name = 
> enterprises.jinny.generic.mmsc.msend-total
> 10:48:56.641 ( 4858) 5 var_jAgentX -> (after) name = 
> enterprises.jinny.generic.mmsc.msend-total.0
> 10:48:56.641 ( 4858) 6 inexact_request_handle -> pass=0
>  
> it thinks that i have sent an exact request.

	No, that's just the search pattern.  Note how it switches to "inexact request" when the exact doesn't match.  Nothing to worry about here...

> Furthermore, 
> below is an extract of the same request when i used a 
> dynamically loaded subagent:

	That's six of one and half a dozen of the other.

> and I think that the subagent agentx should output the same thing

	...to the log?  I don't know why you think this, but I assure you that you are headed down the wrong path with this line of thinking.

> In fact I am fetching the counters from a shared memory, so 
> when the counter isn't found a return NULL should be passed 
> and snmp should go to the next child in the MMSC tree.

	Whoah!

	You haven't provided any details of how you've implemented your MIB, but I think this may be a problem.  Although it is valid SNMP to omit a scalar during a walk, most implementation methods do not support this.  IF "type=6" means "integer", then you should probably return a 0 valued integer instead of null.

> But 
> this is not what's happening. In fact in my shared memory I 
> only have the counters:
>                   JINNY-MMSC-MIB:smsnot-alive(88) type=6
>                   JINNY-MMSC-MIB:smsnot-error(87) type=6
>                   JINNY-MMSC-MIB:smsnot-ok(86) type=6
>                   JINNY-MMSC-MIB:smsnot-total(85) type=6

> and so on. So snmp is walking the tree thousands of times 
> more then he should and i think its because he understood 
> that i sent an exact request.

	I think it's because of a bug in your MIB implementation.  Is this a table or a set of scalars?  Look carefully at what you return when you reach end of MIB.

> So when i use the simple 
> [jinny@florida bin]$ ./snmpwalk -c public localhost mmsc
> Timeout: No Response from localhost
> the request times out and if i use 
> [jinny@florida bin]$ ./snmpwalk -t 10 -c public localhost mmsc
> JINNY-MMSC-MIB::smsnot-total.0 = Counter32: 0
> JINNY-MMSC-MIB::smsnot-ok.0 = Counter32: 0
> JINNY-MMSC-MIB::smsnot-error.0 = Counter32: 0
> JINNY-MMSC-MIB::smsnot-alive.0 = Counter32: 8

	This does not look like "thousands of times".  In fact it looks exactly correct given your statement above of what data you have.  Please clarify.


	HTH,

Mike

-------------------------------------------------------------------------
Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW!
Studies have shown that voting for your favorite open source project,
along with a healthy diet, reduces your potential for chronic lameness
and boredom. Vote Now at http://www.sourceforge.net/community/cca08
_______________________________________________
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