Thanks Bill. I have read (perused) the Redis Sentinel Documentation prior to attempting this and yet again after your comment. I'm trying my best to follow directions yet I'm not getting the expected failover results. I have three RHEL servers with Redis 3.2.1 installed as a service on port 6379 running a replicated group where 143.61.2.31 is the master. The redis-sentinels are likewise set up on port 26379. The new slave config now monitors the master Redis. We have firewalls that I have submitted rules exceptions for ports 6379, 16379 and 26379 bi-directionally between all three servers.
Are there other handshakes going on? I'm watching the debug logs and see what look like pings between the two slaves, but nothing from the master sentinel:
...
6844:X 18 Jul 17:02:01.852 - Accepted
143.61.2.32:580276844:X 18 Jul 17:02:02.886 - Accepted
143.61.2.32:580326844:X 18 Jul 17:02:03.924 - Accepted
143.61.2.32:580376844:X 18 Jul 17:02:04.980 - Accepted
143.61.2.32:580426844:X 18 Jul 17:02:06.061 - Accepted
143.61.2.32:580476844:X 18 Jul 17:02:07.102 - Accepted
143.61.2.32:58052...
28699:X 18 Jul 17:04:27.347 - Accepted
209.82.241.145:5114028699:X 18 Jul 17:04:28.408 - Accepted
209.82.241.145:5114528699:X 18 Jul 17:04:29.421 - Accepted
209.82.241.145:5115028699:X 18 Jul 17:04:30.484 - Accepted
209.82.241.145:5115528699:X 18 Jul 17:04:31.498 - Accepted
209.82.241.145:51160...
I performed a service Redis stop on the Master and both sentinel slaves logged:
6844:X 18 Jul 17:06:02.139 # +sdown master devmaster 143.61.2.31 6379
... and that's as far as it got
The new slave conf file per the comment by Real Bill, follows. It monitors the Master Redis. but the Master Sentinel never finds the slave sentinels. If someone could please identify what I'm missing.