Hi all, I'm using net-snmp 5.2.1 on a MIPS blade running Montavista Linux, and I have a question about the behavior I observed when my net-snmp subagent is talking to a master agent using AgentX. I have to implement an enterprise MIB with 50+ tables, where my subagent is registering only for one row of each table. I'm using the table_dataset helper and the range_subid/range_ubound fields to identify the range for all columns of the row I want to support.
When I then try to walk my MIB, the master agent (which is running on a different server) will try to send first an exact GET using the OID of the first column of the first table row I want to walk, and then ask for a GET_NEXT for each subsequent columns, until the table row is explored, then it moves to the next table, and again only sends my subagent all the OIDs for the single row I'm registered for. This is all working fine when I have rows populated in all tables. But when there are no rows in a given table, even though the table_dataset handler returns NOSUCHINSTANCE, the master agent still keeps asking for all the next columns. Is there a way I can send the master agent a special error code so that it will skip over my entire table row instead of keep asking for the next column anyway? I tried to come up with my handler returning various error, but only NOSUCHNAME seems to stop the master agent, but then it won't jump to the next table and keep the MIB walking anymore. My problem is that in such cases I would like to improve the response time of the subagent. In fact let's assume for a moment I have no table rows populated at all for any table: the master agent will still send the GET for the first column of the first table row I'm registered for, the subagent handler returns NOSUCHINSTANCE, then the master agent sends a GET_NEXT which will cause the snmp subagent to call the handle_getnext_loop() function which in turn will call all 50+ handlers for my tables just to find out there is no next instance of anything. Then the master agent will come down for the same request for the next column which will trigger another huge getnext loop, and so on. So for example if I'm walking an empty table with 30 columns and I have the next 50 tables empty as well, the master agent will call the helper handlers more than 30*50 times just to find out there is no next OID available. Thanks in advance for any suggestion you might have. Regards. Danilo Giulianelli Lucent Technologies Bell Labs Innovations mailto:[EMAIL PROTECTED] ------------------------------------------------------- This SF.Net email is sponsored by: Power Architecture Resource Center: Free content, downloads, discussions, and more. http://solutions.newsforge.com/ibmarch.tmpl _______________________________________________ Net-snmp-users mailing list Net-snmp-users@lists.sourceforge.net Please see the following page to unsubscribe or change other options: https://lists.sourceforge.net/lists/listinfo/net-snmp-users
