Hello, I have a question on agentx and its usage. We are using netsnmp version 5.5 and in our implementation of snmp, we have spawned a separate process, say "trapd" to handle system specific alarms and events. The reasons are two-fold:
1) legacy - The overhead of handling local alarms and events is such an overhead, we didn't want to burden snmpd and I have to use existing legacy code for alert handling from other processes. 2) Desire to keep net-snmp code intact with minimal changes so that in case we need to upgrade, there won't be too much of porting efforts. We have no problem registering for trap sessions and sending traps in v1, v2c and v3 modes and all security levels and have tested successfully with other snmp tools for its correct operation. The problem happens with sending "inform" messages as the INFORM response comes to snmpd and not to our "trapd" process. One idea was to use what "snmptrap" was doing but we cannot wait for timeout period in our daemon(like snmptrap does) as we need to process alerts at the rate of 10 per seconds. I would like your advice/suggestion as to how best to handle this issue: Option 1) Do something similar to sub-agent if that is possible. Option 2) In _sess_process_packet, if we cannot find the session and if it is INFORM response, forward the packet to "trapd". Option 3) ?? I would very appreciate you expert opinion/advice in this matter and any issues you see down the road whatever mechanism we choose. Thanks in advance for your response, Fatima ------------------------------------------------------------------------------ Download Intel® Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev _______________________________________________ Net-snmp-coders mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/net-snmp-coders
