Prime Infrastructure blocks SNMP during sync causing PRTG alarms

[bkoch1]

We're running Prime Infrastructure 1.2, and until yesterday we were mainly using it for wireless monitoring. We use PRTG for monitoring the state of the rest of the network. I added our core router (6509-E w/VSS), and during syncs, we will get PRTG alarms (SNMP traffic sensors) saying "No response". The downtime is ~5 minutes (which is the PRTG scanning interval). So, it seems like Prime is preventing other SNMP traffic during the syncs. Is this normal? I don't really like the thought of receiving a boatload of false alarms several times a day.

[Zoran BKV]

I can confirm this behaviour. Our HP P2000 Controller has stop responding since we started syncing. We just get SNMP error 2003. Also using PRTG to monitor systems.

[Rollin Kibbe]

Hi Brian:

Prime Infrastructure 1.2.0.103 was deferred and removed from Cisco.com in favor of 1.2.1.12, for a number of issues.  The current release is 1.3.0.20 Update 1.  Between the two, there have been a number of fixes with regards to Java threads being locked, amount of system resources being used by the Oracle engine, size of the backups, and a lot more.  You may want to look into getting current to see if the behavior changes.

Clarifying the upgrade path to Prime Infrastructure 1.3.0.20 Update 1

Given the volume of data polled during sync cycles, I could see where Prime Infrastructures questions might dominate or even overwhelm the router's SNMP agent, but there's no process that puts an access-list to prevent the agent processing any other station's traffic while sync is going on.  You may want to check the CPU utilization of the SNMP agent (show process cpu | i SNMP ENGINE) before, during and after the sync to see what impact the sync has.