Hello, were you ever able to resolve this using a keyfile? I’m looking at the same thing. I tried x509 certs briefly, but saw similarly verbose logs. Made me think it was a network-related thing, like a drop and reconnect, but I see no evidence of that in my network logs.
I converted a 5.0.14 stand-alone using authorization to a rs. Used a brain-dead simple configuration; rs config, keyfile on all nodes, bind using DNS name, matching the primary with auth.
Everything is working as expected, but there are endless “Client has attempted to reauthenticate as a single user” logs coming from the IPs of the secondary nodes in ACCESS.
So it is not an issue with the keyfile auth, as the replica set is started, the nodes have initiated, and the data is replicated. The admin DB users are replicating on all nodes as well.
I tried turning off authorization on all nodes and restarting, and it still logs these reauthentication from the nodes to the __system user and local db messages.
There is minimal documentation on the internal SCRAM communication between replica nodes. I looked at the source code, and there is one reference to this log as a warning that the same credentials were reused. So it seems 100% benign in nature. I’d like to figure out what I did wrong, if anything.
{"t":{"$date":"2023-02-27T19:45:33.609+00:00"},"s":"W", "c":"ACCESS", "id":5626700, "ctx":"conn37","msg":"Client has attempted to reauthenticate as a single user","attr":{"user":{"user":"__system","db":"local"}}}
{"t":{"$date":"2023-02-27T19:45:33.609+00:00"},"s":"I", "c":"ACCESS", "id":20250, "ctx":"conn37","msg":"Authentication succeeded","attr":{"mechanism":"SCRAM-SHA-256","speculative":false,"principalName":"__system","authenticationDatabase":"local","remote":"10.2.1.5:58803","extraInfo":{}}}