[QUESTION] Redis 6.2.7: The CLIENT PAUSE <x> WRITE, makes the sentinel failover, expect?

[samof76]

I have a sentinel based setup running with redis 6.2.7.

1 - master
2 - replicas
3 - sentinels

For some testing purpose I would want to do on the current master, to test a client behavior

CLIENT PAUSE 1000 WRITE

But what I have noticed is that the sentinel initiated a failover in sometime, and one of the replicas gets promoted as the slave. Is this an expected behaviour? In which case the question is what is the reasoning behind this?

ALSO: I have not tested this in Redis 7.x

[igxlin]

I cannot reproduce it with the unstable version. It may be fixed.

The PING command sent by sentinel is not a write command. It is not blocked and will not trigger the failover.