No retry logic on poll()/put() for both source/sink
Closed this issue · 1 comments
Currently if Redis isn't available or there is an intermittent network issue both the Sink and Source connector throw errors such as
org.apache.kafka.connect.errors.ConnectException: Could not read messages from stream at com.redis.kafka.connect.source.StreamSourceRecordReader.poll(StreamSourceRecordReader.java:109) at com.redis.kafka.connect.source.RedisSourceTask.poll(RedisSourceTask.java:113) .... Caused by: io.lettuce.core.RedisCommandTimeoutException: Command timed out after 1 minute
Which completely stops the connectors and only a restart of the task will fix it.
For those more experienced with this connector, would setting redis.timeout in the connector config potentially fix this issue? Are there other settings I may have missed?
Does this connector need some retry logic wrapping the poll() and put() function calls (Source/Sink)?