I have two Aruba Mobility Access switch stacks that have recently started the habit of not responding to SNMP get-requests for a period of time each day. It's become predictable - around 12:20 - 12:30 PM each day, our middle-school and high school MDF stacks stop responding to SNMP requests. The middle-school stack will begin responding again within about an hour, the high school stack in about 2.5 hours. The switches continue to function normally during this time. Solar Winds Orion shows them as 'up' because they respond to ping. Only recent change that has occurred that I'm aware of is an upgrade to Solar Winds NCM 7.4. Edge switches in the fabric don't display this behavior; other two MDF stacks (at two elementary schools) MAY, on occasion, show some interfaces in 'warning' state briefly during this time, but come back to normal quickly.
I was thinking that this must be an issue with the devices themselves, but just today I uploaded the most recent firmware to all of them and re-booted all of them and the issue remains the same - the same couple of switch stacks doing the same thing at the same time.
I've run Wireshark traces during the time this happens and it's very obvious that the Solar Winds server is sending out the 'get' requests but no response comes back from the device. Once the 'timer' (whatever it is) has expired and the devices begin responding again, you can clearly see on the traces that they are now sending 'get response' messages back to the Solar Winds server.
It really does appear that this began to happen at the time I upgraded NCM, but I can't see the connection.
Any clues on what might be going on?
