Skip to Content.
Sympa Menu

perfsonar-dev - Re: [pS-dev] "Complex" Request and Reply Examples at SNMP MP

Subject: perfsonar development work

List archive

Re: [pS-dev] "Complex" Request and Reply Examples at SNMP MP


Chronological Thread 
  • From: "Rafael Costa" <>
  • To:
  • Cc: "Loukik Kudarimoti" <>, , Prototype-JRA1 <>
  • Subject: Re: [pS-dev] "Complex" Request and Reply Examples at SNMP MP
  • Date: Wed, 16 Aug 2006 14:34:21 -0300
  • Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:references; b=PM4f2Nqn36VOHGSpFDtFpn/ykFfu/aMqmKat0AHh5dX6yY4qdIeIvurJDEzRtAZANj79tgCvd3b2+jxbYMWp6ODkUISblQUaTetiIGz3PgXXTwybWbVdLpmWevsIB0dUyTcoVITGIRC/xJ4CLXmzGeQ4GeGgQIPdZiq9yGOApGY=

Greetings,

In the first example (Request for key - MetadataKeyRequest) why is there
a datum element in the sent data trigger?  It has been a while since I
have looked at these examples and I am not sure why that is needed in a
key request.

Loukik is the right person to talk about it.

In the NeTraMet examples:

Could you include a little bit more in the descriptions of what each
example would be used for (use cases or something similar)?

I have two kind of requests. The first one is a simple get request (where i provide a OID and get back the data). And the second request type is the one where i provide a OID but i get back a "set" of data (more than one value of a kind). To work with this second one, we "walk" through the MIB and collect the data concerned.

I am still unconvinced to the need for 'type' designations in the
request file (i.e. <parameter name="OID_Alias"
type="numeric">sysUpTime<
/parameter>), but if you are positive you must
do something like this I would suggest it to be:

I agree with you. We are not going to use this extra parameter. When i suggested this one, i was thinking in a better way to describe data (at metadata section), but now i see that it's no more necessary. 

For the ' Request/Response for data request with a "table OID"' example,
the eventType should be used to indicate you are getting a table instead
of just a standard get (SNMP.Table or something similar).  Maybe this is
different in for NeTraMet, but doing 'snmpget ... atTable' fails on my
systems when 'snmptable ... atTable' will succeed.

I agree with you again (i was with this point of view in mind too). We need only to discuss the better name for it.

Thanks for your feedback.

Loukik, we wait for your thoughts.

Regards,

Rafael

On 8/16/06, Jason Zurawski <> wrote:
Rafael & Loukik;

> You can find at
> http://wiki.perfsonar.net/jra1-wiki/index.php/Request_and_Reply_SNMP-MP
> the new request and response examples (as we discussed) about a "more
> complex" get at SNMP MP and NeTraMet Collector integration.
>

In the first example (Request for key - MetadataKeyRequest) why is there
a datum element in the sent data trigger?  It has been a while since I
have looked at these examples and I am not sure why that is needed in a
key request.

In the NeTraMet examples:

Could you include a little bit more in the descriptions of what each
example would be used for (use cases or something similar)?

I am still unconvinced to the need for 'type' designations in the
request file (i.e. <parameter name="OID_Alias"
type="numeric">sysUpTime</parameter>), but if you are positive you must
do something like this I would suggest it to be:

<parameters>
  <parameter name="OID_Alias">sysUpTime</parameter>
  <parameter name="OID_Type">timeticks</parameter>
</parameters>

To avoid having a more complex parameter.

I am guessing that there can be only one item requested per SNMP.Get?
Otherwise it would be impossible to map your requested type parameters
to the datum elements returned?

For the ' Request/Response for data request with a "table OID"' example,
the eventType should be used to indicate you are getting a table instead
of just a standard get ( SNMP.Table or something similar).  Maybe this is
different in for NeTraMet, but doing 'snmpget ... atTable' fails on my
systems when 'snmptable ... atTable' will succeed.

-jason



--
Atenciosamente,

Rafael Costa
NUPERC - UNIFACS
(71) 8805-4840


Archive powered by MHonArc 2.6.16.

Top of Page