Testing fault tolerance

[selcukaltinay]

In order to test fault tolerant hiredis cluster api, I wrote a test code as I mentioned below. At the start of the test I’m starting the 6 redis instance. Then I’m making cluster them. After this process, Im trying to get a key from redis continuously with 10ms delay. Then I’m killing the master nodes and Im expecting that the requests must be answered from replicas(new master). But Im taking core dump. How can I solve this issue without using external retry pattern?

[bjosv]

Interesting, which version of hiredis-cluster are you using?
Could you paste the stack trace from the corefile for us to debug the issue?