Re-connect issues after Redis node restart about hiredis-cluster HOT 5 CLOSED

What is the used connection timeout configuration in hiredis-cluster?
It is controlled by redisClusterSetOptionConnectTimeout()
Without any configuration it might result in long timeouts depending on setup.

from hiredis-cluster.

Hi, we use a timeout of 5 secs.
Yes it is configured with redisClusterSetOptionConnectTimeout()
Just that some connect quickly within seconds & some take up to 5 minutes. All configured with 5 sec connection timeout.

from hiredis-cluster.

Sounds good with the configured timeout.

If I understand correctly you are running in Kubernetes, and I guess you have tried to manually setup a connection from the hiredis-cluster pod to the Redis pod? Like using telnet just to make sure the connection should work.
Another option would be to increase the logging in the redis node which would tell if redis sees the connection attempt.

Another area that might affect is the eventsystem that is needed for async.
Are you using an own implementation, or a public one that we have tested or can use to try to reproduce the problem?
When hiredis attempts to connect to a node it registers the socket to the eventsystem, which poll the socket for a write event or to have it timeout as in your case. This might affect the results as well.

from hiredis-cluster.

Our adapter is based on the reference adapter. It is similar to the ones in your repo and is well tested.

Ok, we will check if there could be any issue with the connectivity to this Redis node. Ideally there shouldn't be, as the other clients running in the same pod are able to re-connect immediately.

from hiredis-cluster.

This is very likely fixed by #122, please reopen if not.

from hiredis-cluster.