I have a redis sentinel structure with 3 redis instances and 3 redis-sentinel instances. When I make them up for the first time, slaves constantly changes their master to 127.0.0.1 and then back to real master and then to 127.0.0.1 and so on. When I force the master to be down for 20 sec, another instance becomes master and everything becomes stable. Here are my config. What can cause this and how can I solve this problem? Logs are constantly saying fix-slave-config when sentinels are unstable.
initial master config
daemonize yes
port 6380
bind 0.0.0.0
supervised systemd
pidfile "/run/redis/redis-sentinel.log"
logfile "/var/log/redis/sentinel.log"
sentinel monitor mymaster 127.0.0.1 6379 2
sentinel auth-pass mymaster ***
sentinel down-after-milliseconds mymaster 4000
sentinel failover-timeout mymaster 2000
sentinel parallel-syncs mymaster 1
sentinel announce-ip IP_OF_CURRENT_MACHINE
initial slaves' config
daemonize yes
port 6380
bind 0.0.0.0
supervised systemd
pidfile "/run/redis/redis-sentinel.log"
logfile "/var/log/redis/sentinel.log"
sentinel monitor mymaster MASTER_IP 6379 2
sentinel auth-pass mymaster ***
sentinel down-after-milliseconds mymaster 4000
sentinel failover-timeout mymaster 2000
sentinel parallel-syncs mymaster 1
sentinel announce-ip IP_OF_CURRENT_MACHINE
What I tried:
quorum= 1,2
down-after-milliseconds 4000,10000
failover-timeout 2000,6000
