Configure the Cassandra Nodes

Last updated March 2020

The Cassandra distributed database can be configured to run on its own cluster of nodes OR together on the MTA nodes. You must know the name and IP address of each Cassandra cluster node (or Platform node if you are combining it with an MTA) before you can configure Cassandra.

Follow all steps below on the first Cassandra node until reaching the final numbered step in this section, which is to configure the remaining nodes.

Each Cassandra cluster must have a unique name that is shared by all nodes in that cluster. The name of the cluster is used to prevent machines in one logical cluster from joining another. All nodes participating in a cluster must have the same cluster_name value. Create a Cassandra cluster name and set it as an environment variable.
```
export CLUSTER="CAS-12345678"
echo $CLUSTER >/var/tmp/cass.cluster
```
Copy the /opt/msys/3rdParty/cassandra/conf/cassandra.yaml.dist file to /opt/msys/3rdParty/cassandra/conf/cassandra.yaml.
Edit the /opt/msys/3rdParty/cassandra/conf/cassandra.yaml file. Set or change the following configuration options:
- cluster_name: 'your-cluster-name '
- listen_address: your-firstCassNode-IP
- rpc_address: 0.0.0.0
- rpc_max_threads: 200
- thrift_framed_transport_size_in_mb: 25
- Replace the default Cassandra root /var/lib/cassandra/ with /var/db/cassandra/ in all three places it appears (data_file_directories, commitlog_directory, and saved_caches_directory). Be sure to retain the existing directory names beneath the root at each of those places.
- Ensure the following option is formatted as shown below:
```
seed_provider:
- class_name: org.apache.cassandra.locator.SimpleSeedProvider
  parameters:
  - seeds: "
```
  cassnode1.example.com,cassnode2.example.com,cassnode3.example.com"
  
  Configure EXACTLY THREE Cassandra hostnames in the seeds parameter. It is not critical which of your Cassandra hostnames are in this list as long as the parameter has exactly three hostnames set.
  
  Note
  
  If you are performing a single node install, configure one Cassandra hostname in the seeds parameter.

Set your environment variables to handle singlenode OR multinode configurations.

For environments with a single Cassandra node, run the following:

export CASS_MULTI_OR_SINGLE=singlenode
export CASS_CLUSTER_OR_SINGLE=singlenode

For environments with multiple Cassandra nodes, run the following:

export CASS_MULTI_OR_SINGLE=multinode
export CASS_CLUSTER_OR_SINGLE=cluster

Start the Cassandra service.
```
service msys-cassandra start
sleep 30
```
Warning

It is possible to get a FAILURE result message on first startup, even though the startup actually succeeded. Perform the next step below to determine the real status. If you do not get the indicated result, check logs at /var/log/msys-cassandra/ for error messages.

Check the status of the Cassandra service.

service msys-cassandra status

Your results should look something like this:

Datacenter: datacenter1
=======================
Status=Up/Down
|/ State=Normal/Leaving/Joining/Moving
--  Address      Load       Tokens  Owns (effective)  Host ID                               Rack
UN  10.77.0.158  45.89 KB   256     100.0%            a5732aca-b808-417b-a83b-ac97b4693694  rack1

Apply the initial Cassandra schema. Do this only on the first Cassandra node in the cluster.

Note

You can safely ignore any 'Invalid column name' messages that may appear.

export PATH=/opt/msys/3rdParty/bin:$PATH export IPADDRESS=host $HOSTNAME | cut -d' ' -f4

cd /opt/msys/ecelerity/etc

../bin/cassandra_momo_setup.sh --$CASS_MULTI_OR_SINGLE $PWD |
tee -a /opt/msys/app/users-api/cql/cassandra_schema.log

cd /opt/msys/app/users-api/cql

cqlsh $IPADDRESS -f load_${CASS_CLUSTER_OR_SINGLE}_keyspace.cql
2>&1 | tee -a /opt/msys/app/users-api/cql/cassandra_schema.log

cqlsh $IPADDRESS -k authentication
-f upgrades/V2015.01.20_00.00.00__create_customer_metadata.cql
2>&1 | tee -a /opt/msys/app/users-api/cql/cassandra_schema.log

cqlsh $IPADDRESS -k authentication
-f upgrades/V2015.01.21_00.00.00__add_terms_of_use_column.cql
2>&1 | tee -a /opt/msys/app/users-api/cql/cassandra_schema.log

cqlsh $IPADDRESS -k authentication
-f upgrades/V2015.05.21_00.00.00__add_email_verified_column.cql
2>&1 | tee -a /opt/msys/app/users-api/cql/cassandra_schema.log

cd /opt/msys/app/webhooks-api/cql

cqlsh $IPADDRESS -f load_${CASS_CLUSTER_OR_SINGLE}_keyspace.cql
2>&1 | tee -a /opt/msys/app/users-api/cql/cassandra_schema.log

cqlsh $IPADDRESS -k webhooks
-f upgrades/V2015.08.12_12.00.00__add_auth_columns.cql
2>&1 | tee -a /opt/msys/app/users-api/cql/cassandra_schema.log

Configure the remaining Cassandra nodes

Configure Cassandra's yaml file on all remaining nodes.

From the first Cassandra node, copy /opt/msys/3rdParty/cassandra/conf/cassandra.yaml to each remaining Cassandra node.

cd /opt/msys/3rdParty/cassandra/conf
scp cassandra.yaml cass2.com:/opt/msys/3rdParty/cassandra/conf/

On every node, edit /opt/msys/3rdParty/cassandra/conf/cassandra.yaml to be sure that this option listen_address: w.x.y.z matches the IP address of the node.
Start the Cassandra service on each node.

Warning

Do not start the Cassandra on two or more nodes simultaneously. Allow Cassandra startup to fully complete on an individual node prior to starting on the next one.

service msys-cassandra start

After you have started Cassandra service on each node, check the status on any node.

service msys-cassandra status

You should see all the nodes in the cluster. Something like this:

Datacenter: datacenter1
=======================
Status=Up/Down
|/ State=Normal/Leaving/Joining/Moving
--  Address      Load       Tokens  Owns   Host ID                               Rack
UN  10.77.0.177  91.92 KB   256     31.2%  fbe23d7b-0c7b-4a45-90bb-20373ff3948b  rack1
UN  10.77.0.233  110.98 KB  256     35.4%  2713fb8f-b188-4b0b-9d11-b76506e922bc  rack1
UN  10.77.0.235  111.41 KB  256     33.3%  88a7616c-36b9-4aa0-9ddc-db24ad76115c  rack1

Was this page helpful?

Configure the Cassandra Nodes

Note#

Warning#

Note#

Warning#

Note

Warning

Note

Warning