Hi. I'm using Net-SNMP Version: 5.7.3. My sub-agent is running into a network device used in fiber optic networks. The device is an OLT. The sub-agent is implemented in the ZebOS operation system (Linux zebos 3.16.7). This sub-agent uses the AgentX implemented in the ZebOS too. The Agent comes from the NET-SNMP.
I can send SNMP Gets to the Agent and receive the answers in the Manager normally. Only AgentX is being used here. Whether I run a SNMP Walk, when the agent is undisturbed for a while, everything is ok too. However, when my device is generating too many traps (this is part of a stress test and so such test creates all traps), any SNMP started Walk in the manager doesn't get the answers with a satisfactory performance. That is, I see the manager presenting the answers , one by one, slowly. Could it be a case of agent performance? Or could it be a case of agent configuration? Is there some kind of prioritization between gets and traps, considering the agent logic? If yes, how to configure it and take control of the priority for each kind of both SNMP message? P.S.: If I comment the line, in the AgentX source code, that sends the trap PDU to the Agent, leaving all the rest executing normally, the SNMP Walk becomes fast again. So, it seems that the Agent slows the SNMP GETs execution, when it is handling may traps. Any hint will be very helpful! Best regards! RODRIGO PIMENTA CARVALHO Inatel Competence Center Software Ph: +55 35 3471 9200 RAMAL: 979 _______________________________________________ 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
