renewClusterSlots throw JedisException

[jgaoIXL]

Expected behavior

when calling ClusterConnectionProvider.renewSlotCache(). it should succeed whether there are dead nodes in the cluster or not.

Actual behavior

When I run the jedis client against a auto scaling redis cluster where the nodes are constantly going in and out, ClusterConnectionProvider.renewSlotCache() throw JedisException, the stack traces are like below. So seems to me that there are dead nodes in the cluster that cause the getResource() to throw JedisException, however, shouldn't that exception to be caught?

redis.clients.jedis.exceptions.JedisException: Could not get a resource from the pool at redis.clients.jedis.util.Pool.getResource(Pool.java:38) at redis.clients.jedis.ConnectionPool.getResource(ConnectionPool.java:28) at redis.clients.jedis.JedisClusterInfoCache.renewClusterSlots(JedisClusterInfoCache.java:123) at redis.clients.jedis.providers.ClusterConnectionProvider.renewSlotCache(ClusterConnectionProvider.java:58)

Steps to reproduce:

Recreating this will be tough, basically what I was doing is run the Jedis against an auto scaling redis cluster. And this exception happen after a while. Anyway, shouldn't JedisException be caught inside renewSlotCache()? Like having some dead nodes should be tolerated.

Redis / Jedis Configuration

Jedis version:

4.2.3

Redis version:

6.2.2

Java version:

17

[sazzad16]

@jgaoIXL Thank you for reporting this. And sorry to miss it before releasing 4.3.0.