Hi Dojo, a simple question:
When using the redundant polling feature (with SNMP), will DataMiner already switch to using the alternate IP as soon as a single timeout occurs on a parameter OID request timing out?
OR, will the element first perform the "RETRIES" that are configured on the element settings before making that switch?
What about multiple parameters in one group (without the MULTIPLEGET option), will it try to complete the group on the connection that is now failing? or will it have switch to the redundant connection?
The docs are not clear on this: Redundant polling | DataMiner Docs
Thank you.
From what I can see from a quick test, it will execute all retries, then switch to the redundant connection while also moving to the the next poll group, so it will not retry the same group again on the new connection.
I couldn't test specifically with a singular timeout, but if a single param is failing in a group without multipleGet, it will do the normal behavior and keep trying, before moving on to the rest of the group, so the redundancy is not triggered if the group is fine otherwise
Thank you! This seems correct, and in the meantime the documentation also got expanded on this topic. Thank you for that!
https://docs.dataminer.services/develop/devguide/Connector/ConnectionsRedundantPolling.html#scenarios-that-trigger-redundant-polling-switching
To add on to this, if polling for an SNMP parameter results in an error like NO SUCH NAME, the element would also switch polling interfaces.
If the element is polling an invalid SNMP parameter, the element would switch back and forth between interfaces based on the frequency that this parameter is being polled.
This case is unlikely to happen, but this is extra information on its behaviour.