When I first boot the NioNode I am able to query the SNMP agent (for example SNMP GET to get the name of the currently deployed project). After an hour or two though, I am unable to get a response to an SNMP GET. If I re-boot from the web interface, I again have no problems.
I have tried turning off and then on the SNMP functionality from the web interface. I also tried exposing an SNMP control inthe hope that this would keep the agent alive, but no luck.
I am using HP Openview and the internal MIB Browser provided with Openview.
Thanks for any input!

I am experiencing this issue as well as of version 1.6.5. The NioNode stops responding to SNMP all together after an hour or two of running. It would appear that this only happens once SNMP has been queried. In other words, the SNMP daemon will work fine even days after the NioNode boots, but once I hit it with a few SNMP GETs, it's just a matter of time before the SNMP daemon dies.
This makes it really difficult to monitor our collection of Nions in the field. Not being able to monitor temperature, audio status, etc. is a real hindrance.
Has anyone found any workarounds or solutions to this problem?
Help!

Upon further research, here are some things I've learned:
1) On the Nion, snmpd is munching CPU, stuck at 75-90% CPU. I suspect this is some kind of I/O wait state that's blocking
2) snmpd cannot be restarted. It is unresponsive to SIGTERM
This could be a bug in the version of snmpd that is used on the Nion, or maybe it is stuck trying to read from one of the other Nion processes? No idea ...

Thanks for the report.  We are investigating.

I need a bit more info here.  I've deployed a simple project to a NION and am polling the project name every 5 seconds through SNMP.   So far (after three hours) it is working fine, top shows no unusual CPU usage.  Any other ideas on what I should try?

cwa wrote:
I need a bit more info here.  I've deployed a simple project to a NION and am polling the project name every 5 seconds through SNMP.   So far (after three hours) it is working fine, top shows no unusual CPU usage.  Any other ideas on what I should try?
Hmmm not sure what the difference could be.
I'll do some more testing when I'm back in the office on Monday. Could you post your simple program? I could compare it our monolithic program and see if I notice any significant differences.

Project is simply the NION device, nothing else.  I left it running overnight, all was well.  Thanks for following up!