Configuring SRM Driver retry behaviour

If a target Kafka cluster is not available at startup, the replications targeting that cluster are not set up. In a case like this, the Streams Replication Manager (SRM) Driver sets up the replications for available clusters and retries replication setup for the unavailable clusters. Retry behaviour can be configured in Cloudera Manager.

By default the SRM Driver is configured to retry indefinitely with a retry interval of 10000 ms (10 seconds). Both the retry count and retry interval can be configured in Cloudera Manager using the Retry Count for SRM Driver and Retry Delay for SRM Driver properties.

Retry Count for SRM Driver
Configures the maximum number of retries. The default value of -1 means that retries are carried out indefinitely. If the configured retry count is reached, the SRM Driver will stop. Additionally, setting the property to 0 is the same as setting it to 1.
Retry Delay for SRM Driver
Configures the interval at which retries happen, this is the delay (in ms) between each retry.
  1. In Cloudera Manager, select the Kafka service.
  2. Go to Configuration.
  3. Find and configure the following properties:
    • Retry Count for SRM Driver
    • Retry Delay for SRM Driver
  4. Click Save Changes.
  5. Restart the Streams Replication Manager service.
Retry behaviour is configured for the SRM Driver. The SRM Driver uses the configured values if a Kafka cluster is unavailable at startup.