We recently observed an issue in which a Redis Enterprise cluster running RedisGraph received an unhandled query that caused a server crash, and redis-cli re-emitted the query to the promoted slave, which also crashed.
The issue is not reproducible when the command is emitted via bdb-cli or telnet, so I am fairly confident that it is triggered by redis-cli.
What retry mechanisms are built into redis-cli, and how can we prevent failing commands from being retried?
Comment From: oranagra
I think I was actually going to look into that mechanism in #8930. I'll consider this scenario too and test it.