Re: dbsnmp problem

"N. Gatt" <ngatt_at_yahoo.com>
??????:947486db.0403261525.706ed053_at_posting.google.com...
> Hi
>
> I'm running dbsnmp for OPS 8.1.7 on AIX 4.3 as a peer agent to the
> native snmpd. Everything works fine except when I do an snmp walk
> (get-next) on the Oracle MIB tree (.1.3.6.1.4.1.111). This only
> produces an output for the oraDbSysEntry sub-tree
> (.1.3.6.1.4.1.111.4.1.1.1). However, when I do a walk on each of the
> sub-trees individually, I get the full set of OIDs for each sub-tree.
> It's only when the snmp walk is done from the top of the Oracle MIB
> tree, that things go wrong.
>
> I've traced the master agent (snmpd), which produced the following
> output (dbsnmp responce) when the get-next stops traversing the whole
> tree :
>
> get-response {
> request-id 84,
> error-status 5,
> error-index 1,
> variable-bindings {
> VarBind {
> name 1.3.6.1.4.1.111.4.1.1.1.24.5,
> value {
> empty NULL
>
>
> Any clues on why this haappening will be greatly appreciated.
>
> Thank You
>
> Noel

Hi Noel,

I think the system is ok and the get-next still works. The snmp walk should be within the MIB tree. How's the tree root defined in the system? I think it should be on the node --- Oracle MIB tree (.1.3.6.1.4.1.111). Above that node, "the top of the Oracle MIB tree" you said, is just looked like a tree and Oracle MIB tree root hooked, such as (.1.3) or (.1.3.6). But actually it is not defined in the agent. Otherwise, the root should not be on the node (.1.3.6.1.4.1.111). Just find out the tree root regardless how the node is written. The nodes contained inside the root node are virtual.

Furthermore, get-next command stops when it can not find further more node else, or when the Agent Device is too busy to process the request. And finally, the agent always sends a get-response with null to console to end get-next command. You snmp walk outside the MIB tree defined in the agent, of course you will get 'NULL'.

Johnny Jiang Received on Sat Mar 27 2004 - 23:58:25 CET