This is a discussion on Re: Help with snmp tables - SNMP ; On 25/03/2008, Baltazar Francois wrote: > We have generate the C code from the mib file OAM-GLOBAL-REG-V1.txt > (attached) The first thing that strikes me about that MIB file is that you are defining a table with no accessible content. ...
On 25/03/2008, Baltazar Francois
> We have generate the C code from the mib file OAM-GLOBAL-REG-V1.txt
The first thing that strikes me about that MIB file is that you are
defining a table with no accessible content.
The oamIndex column is defined as being "not-accessible"
(which is correct), but the two "column1/2" objects are also
defined as "not-accessible".
This means that there is nothing in this table which can
be retrieved via SNMP.
> with the command
> mib2c -c mib2c.create-dataset.conf oam
I'm not convinced that this code is the output of mib2c.
Given that all three MIB objects are not-accessible,
then I'd expect the default row call to be empty.
The OID given in 'oamTable_oid' doesn't match the OID from the
MIB file either.
> Then we have created the handler to respond to the mib values.
If you are using the 'table_dataset' helper, then you don't need
a handler routine. The dataset handler will deal with this for you.
> When we have test the code , the agent seem to respond only to GET NEXT
> request. The OID that results from these requests is something like:
What is the exact GET request(s) that you are trying?
> As you can see in the .c attached file there is a call to
> netsnmp_table_row_add_index(row, ASN_OCTET_STR, "snmpv3",
Again, I don't believe that this was generated by mib2c.
> We have tried to change this call to replace by an integer value (the value of
> the index column),
That would be correct, yes.
> code has compiled but crash at this point giving "segmentation fault".
What is the exact code that you are using?
> What we are trying to do is only to have an only one handler for all the values
> of the table, which respond to GET and GETNEXT requests for the values
> which look like column_oid.1, column_oid.2, column_oid.n....
That is how things should work, yes.
Apart from the problems mentioned above (like not having any valid
columns in your MIB!)
> Note : We are using net-snmp version 5.1.1 on Red Hat EL4
That is a very old release, and the 5.1.x line is no longer supported.
You might wish to upgrade to a more recent version (e.g. 5.4.1).
Though the problems you have outlined are probably not version
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
Net-snmp-coders mailing list