The error message below was asked several hours ago on the Telegram group with no reply yet. We ourselves began seeing this message in relay logs immediately after the hard fork:
$ grep HardForkEncoderDisabledEra node.log | head -1
[relay:cardano.node.ErrorPolicy:Error:84] [2020-12-16 21:45:37.94 UTC] [String "ErrorPolicyUnhandledApplicationException (HardForkEncoderDisabledEra (SingleEraInfo {singleEraName = \"Allegra\"}))",String "ErrorPolicyTrace",String "207.244.225.1:41737"]
I’m going to venture a guess, and assume unless someone posts otherwise, that this is the result of incoming peers connecting to our 1.24.2 relays who are still on a previous version of the node not compliant with the Allegra hard fork. It would be reassuring to get insider confirmation of this & maybe a developer explanation in case it’s not that simple. 
BTW if this suspicion is right then there may be a consistency problem since we have a lot of these messages: not yet counted individual IPs, but in all about 9 every minute since the beginning of the epoch 9 hours ago:
$ grep HardForkEncoderDisabledEra node.log | wc -l
4683
I would recommend to submit a ticket in github as well:
I’d like to keep this issue alive, since a link to the Github repository we’re all already using doesn’t address the original question, which could still benefit from some high level discussion.
There was since a confirmation yesterday from @rdlrt on the Telegram group confirming that the message above in fact is from nodes with the pre-Allegra software:
I have also seen this error a few times. But since this morning it’s only the usual “Failed to start all required subscriptions” once in a while. So I guess, I actually do not any peer remaining on pre-1.24.2 
1 Like
There was since a confirmation yesterday from @rdlrt on the Telegram group confirming that the message above in fact is from nodes with the pre-Allegra software
Same errors now happening after Alonzo Hard Fork on my relay nodes:
ErrorPolicyUnhandledApplicationException (HardForkEncoderDisabledEra (SingleEraInfo {singleEraName = “Alonzo”}))
I was surprised that I had to replace 25% of the entries in mainnet-topology.json to only have Alonzo-compatible outgoing peers in there. After node restart there is only one of these errors left, of an incoming peer.
yes, on each of our relays we had to add one or two IP addresses to an inner firewall ufw to refuse connections from pre-Alonzo hosts who were still choosing us as a peer after the fork. Not having these incoming connections from the obsolete nodes seems to make a big improvement in performance.
1 Like