Configuration Overview

Learn how to configure SRM with the srm.properties configuration file.

SRM configuration is managed through a single configuration file. The default location of the configuration file for .rpm based installs is /opt/streams-replication-manager/config. For ZIP or TAR based installations the default location is $SRM_CONF_DIR/srm.properties. The default name of the file is srm.properties.

Command line tools provided with SRM read configuration properties from the configuration file. All tools automatically collect configuration information from the default file. If required it is possible to store the configuration file at a different location. In this case however, you need to specify the configuration file and its location when using the tools. Specifying an alternate location for the file is achieved with the --config option which is supported by all SRM command line tools.

The srm.properties file accepts all Kafka client properties available in the version of Kafka that you are using. Additionally, it supports a number of SRM specific configuration properties. For a comprehensive list of SRM specific properties, see Configuration Properties Reference.

Configuration properties for SRM can be set on three different levels, which are the following:

Top level: Top level or global configuration is achieved by adding the property on its own. For example:
```
replication.factor=3
```
Cluster level: Cluster level configuration can be achieved by prepeding the configuration property with a cluster alias. For example:
```
primary.replication.factor=3
```
Replication level: Replication level configuration can be achieved by prepeding the configuration property with the name of the replication. For example:
```
primary->secondary.replication.factor=3
```

Minimum Configuration

At minimum the configuration file has to contain cluster aliases, cluster connection information, and at least one cluster->cluster replication that is enabled.

Cluster aliases are specified with the clusters property. Aliases are arbitrary names defined by the user. They are used in other configuration properties as well as with the SRM command line tools to refer to the clusters added for replication. For example:

#Kafka cluster aliases
clusters = primary, backup

Each cluster alias has to have connection information associated with it. Connection information is specified with the bootstrap.servers property. You add connection information by prepending the bootstrap.servers property with a cluster alias and adding the address of the Kafka broker as the value. When configuring connection information, add each cluster to a new line. If a cluster has multiple hosts, add them to the same line but delimit them with commas. For example:

#Kafka broker addresses
primary.bootstrap.servers = primary-cluster1.vpc.example.com:9092, primary-cluster2.vpc.example.com:9092, ...
backup.bootstrap.servers = backup-cluster1.vpc.example.com:9092, backup-cluster1.vpc.example.com:9092, ...

Cluster replications can be set up and enabled as follows:

primary->backup.enabled = true
backup->primary.enabled = true

The default srm.properties shipped with SRM contains examples for cluster aliases, connection information, and cluster->cluster replications. In addition, it also contains a number of other pre-set properties. These properties however are not required for SRM to function, they are included to give users a basic example of how the configuration file can be set up. Cloudera recommends that you study the example configuration provided in Configuration Examples to find out more about how SRM should be set up for different replication scenarios.