Hi ! I have set up a sentinel architecture with 3 nodes and 3 sentinels. I have activated tls on each type (node and sentinels) When I consult the sentinel.conf I can see this :

# Generated by CONFIG REWRITE
sentinel myid myid1
sentinel config-epoch redis-primary 2
sentinel leader-epoch redis-primary 2
sentinel known-replica redis-primary hostname1 6379
sentinel known-replica redis-primary hostname2 6379
sentinel known-sentinel redis-primary hostname3 26379 myid2
sentinel known-sentinel redis-primary hostname2  26379 myid3
sentinel known-sentinel redis-primary hostname3  **0** myid4

Why is this hostname3 0 knowned by the sentinel as it should not be used in tls ?

I can provide the config but I prefer to not post it here, by email would be preferable.

Comment From: hwware

@mmorel-35 Could you please send your detail profile to my github email? I could help check this issue. Thanks

Comment From: sjpotter

As an FYI, I recently ran into a similar/same issue, when I converted an existing small test cluster from running in plain text to running just over tls.

It's possible it was because I was converting nodes 1 by 1, that there's a bug that caused it to record a node on port 0 in the process. When I wiped out the logged configuration from my sentinel.conf files and restarted all the sentinels, it came up without the phantom node on port 0.