Starting DataStax Enterprise as a service

Steps for starting the DataStax Enterprise (DSE) service when DataStax Enterprise was installed from RHEL or Debian packages.

All nodes types are DataStax Enterprise nodes and run the database.

Considerations for starting a cluster

Be aware of the following when starting a DataStax Enterprise cluster:

Nodes must be segregated by datacenters

Transactional, DSE Search, DSE Analytics, and SearchAnalytics nodes must be in separate datacenters. For example, in a cluster with both DSE Search and transactional nodes, all DSE Search nodes must be in a one or more search datacenters and all transactional nodes must be in one or more datacenters.

DSE Graph can be enabled on any node in any datacenter.

Deploying a mixed-workload cluster

When deploying one or more datacenters for each type of node, first determine which nodes to start as transactional, analytic, DSE Graph only, DSE Graph plus other types, DSE Search, and SearchAnalytics nodes. Deploy in this order:

  1. Analytic seed nodes.

  2. Transactional or DSE Graph-only seed nodes.

  3. DSE Search seed nodes.

  4. SearchAnalytics nodes.

  5. Remaining nodes one at a time. See Initializing multiple datacenters per workload type.

DSE Analytics nodes

Before starting DSE Analytics nodes, ensure that the replication factor is configured correctly for the analytics keyspaces. Every time you add a new datacenter, you must manually increase the replication factor of the dse_leases keyspace for the new DSE Analytics datacenter.

Start up commands

Set the type of node in the /etc/default/dse file. (Start-up scripts are also available in /etc/init.d.)

Command Description


Starts the node as a DSE Graph node.


Starts the node as a Spark node and starts the Spark Master service.


Starts the node as a DSE Search node.

Transactional-only or BYOS nodes

<NODE_TYPES>=0 or not present.

Node type Settings

Spark Analytics node




No entry is the same as disabling it.

Spark Analytics, DSE Graph, and DSE Search node


BYOS (Bring Your Own Spark)

Spark nodes run in separate Spark cluster from a vendor other than DataStax.

Set BYOS nodes as transactional nodes:

All_NODE_TYPES=0 or not present.

DSE Graph and BYOS


An integrated DSE SearchAnalytics cluster allows analytics jobs to be performed using CQL queries.



  1. If DataStax Enterprise is running, stop the node.

  2. Set the node type in the /etc/default/dse file. For example, to a Spark node:


    Alternately, you can omit the other start up entries and just use SPARK_ENABLED=1.

  3. Start DataStax Enterprise:

    sudo service dse start

    If the following error appears, look for DataStax Enterprise times out when starting and other articles in the Support Knowledge Center.

    WARNING: Timed out while waiting for DSE to start.
  4. To verify that the cluster is running:

    nodetool status

    If DSE has problems starting, look for starting DSE troubleshooting and other articles in the Support Knowledge Center.

    The nodetool command shows the node type and the status. For a transactional node running in a normal state (UN) with virtual nodes (vnodes) enabled shows:

    Datacenter: Cassandra
    |/ State=Normal/Leaving/Joining/Moving
    --  Address    Load       Tokens  Owns    Host ID                               Rack
    UN  82.43 KB   128     ?       40725dc8-7843-43ae-9c98-7c532b1f517e  rack1

    For example, a running node in a normal state (UN) with DSE Analytics without vnodes enabled shows:

    Datacenter: Analytics
    |/ State=Normal/Leaving/Joining/Moving
    --  Address         Load       Owns    Host ID                               Token                 Rack
    UN  103.24 KB  ?       3c1d0657-0990-4f78-a3c0-3e0c37fc3a06  1647352612226902707   rack1

Was this helpful?

Give Feedback

How can we improve the documentation?

© 2024 DataStax | Privacy policy | Terms of use

Apache, Apache Cassandra, Cassandra, Apache Tomcat, Tomcat, Apache Lucene, Apache Solr, Apache Hadoop, Hadoop, Apache Pulsar, Pulsar, Apache Spark, Spark, Apache TinkerPop, TinkerPop, Apache Kafka and Kafka are either registered trademarks or trademarks of the Apache Software Foundation or its subsidiaries in Canada, the United States and/or other countries. Kubernetes is the registered trademark of the Linux Foundation.

General Inquiries: +1 (650) 389-6000,