Thanks Dave for the detailed explanation.
I overlooked the limitation of snmpwalk even after Valantina has given
the example.


-----Original Message-----
From: [] On
Behalf Of Dave Shield
Sent: Thursday, May 08, 2008 3:40 PM
To: Pamidipati Suresh-G20238
Cc: valantina arumugam;
Subject: Re: How to differentiate snmpwalk & snmpgetnext request ??

2008/5/8 Pamidipati Suresh-G20238
> Regarding the example you mentioned, why do you want to gather all the

> instances of the table in the 1st getnext itself when snmpwalk is
> given.?

I believe that Valantina is concerned about the data changing while the
walk is taking place.

For example, consider a table containing two columns and (initially) two
columns. Walking this table would look like:

GETNEXT table --> this.1
GETNEXT this.1 --> this.2
GETNEXT this.2 --> that.1[*]
GETNEXT that.1 --> that.2
GETNEXT that.2 --> endOfTable

The client could then display the two rows quite happily:

this.1 that.1
this.2 that.2

But suppose that a new row (3) was created at exact moment that the walk
had got to the point marked[*] The remaining sequence of GETNEXT
requests would then be

GETNEXT that.1 --> that.2
GETNEXT that.2 --> that.3
GETNEXT that.3 --> endOfTable

leaving the client with the inconsistent table

this.1 that.1
this.2 that.2
(?) that.3

Alternatively, suppose the second row was deleted at[*]
The walk would then finish

GETNEXT that.1 --> endOfTable

giving the client a table:

this.1 that.1
this.2 (?)

Loading all the data at the start of the walk would allow the client to
retrieve a consistent set of information (albeit one that was slightly
out of date).


__________________________________________________ ____________________
This email has been scanned by the MessageLabs Email Security System.
For more information please visit
__________________________________________________ ____________________

This email is sponsored by the 2008 JavaOne(SM) Conference
Don't miss this year's exciting event. There's still time to save $100.
Use priority code J8TL2D2.;
Net-snmp-users mailing list
Please see the following page to unsubscribe or change other options: