Recently, we met a strange error in our prod. We have a REDIS cluster with 3 masters and 3 slaves. 77:7000, 99:7001, 13:7002 are the masters. While I use redis-cli to connect 77:7000 and execute "cluster nodes" command. The info is like below:
It seems like 77:7000 is importing lots of slots out of its range. And output of "cluster info" is like:
While from 99:7001, 13:7002, the output of "cluster info" are all ok. When we want to query keys from 77:7000, it will show me that the cluster is down. From the 99:7001, 13:7002, their keys can be queried.
My solution is to failover 77:7000 to its slave manually, then everything is ok. After this, I failbacked to 77:7000 and execute "cluster info", the output is cluster ok.
But any idea about the weird problem? Thanks a lot!
