Describe the bug

Hi guys. I have a 3-node cluster with TLS replication and which appear to work fine and logs are apparently error-free but as soon as add 'sentinel' logs get lots of: ... Error accepting a client connection: error:1408F10B:SSL routines:ssl3_get_record:wrong version number (conn: fd=9) ... Remove "sentinel" from the equation, just stop the service, and errors disappear.

I'm on CentOS Stream with redis-6.0.9

many thanks,L

A short description of the bug.

To reproduce

Steps to reproduce the behavior and/or a minimal code sample.

Expected behavior

A description of what you expected to happen.

Additional information

Any additional information that is relevant to the problem.

Comment From: hwware

Hello @lejeczek , I think you also need to correctly setup the sentinel with tls mode. Also I noticed you already used cluster mode, so what is the purpose you using sentinel in this case? Can you give some more context here? thanks!

Comment From: lejeczek

Not that Redis' official docs mention anything specific - actually opposite - if I remember correctly they say 'sentinel' for TLS will use redis conf for TLS too. I only follow those docs by Redis, which should be easy to reproduce. Sentinel is needed for HA, as I understand you cannot have it any other way. many thanks, L.

Comment From: binocvlar

@lejeczek :

if I remember correctly they say 'sentinel' for TLS will use redis conf for TLS too

The docs state the following:

Sentinel inherits its networking configuration from the common Redis configuration, so all of the above applies to Sentinel as well. When connecting to master servers, Sentinel will use the tls-replication directive to determine if a TLS or non-TLS connection is required.

This wording is confusing, and I was mislead by this as well. I interpreted this to mean that the sentinel process will inherit settings from the Redis configuration, but I believe what the author actually meant is that sentinel uses configuration parsing code that is shared in common with redis-server.