Cassandra

This section is about how to perform a specific task. If you want to understand how a certain component works, please see Reference

The rest of the page assumes you installed using the ansible playbooks from wire-server-deploy

For any command below, first ssh into the server:

ssh <name or IP of the VM>

This section only covers the bare minimum, for more information, see the cassandra documentation

Check the health of a Cassandra node

To check the health of a Cassandra node, run the following command:

ssh <ip of cassandra node> /opt/cassandra/bin/nodetool status

or if you are running a newer version of wire-server (altough it should be backwards compatibile)

ssh <ip of cassandra node> /opt/cassandra/bin/nodetool -h ::FFFF:127.0.0.1 status

You should see a list of nodes like this:

Datacenter: datacenter1
=======================
Status=Up/Down
|/ State=Normal/Leaving/Joining/Moving
--  Address         Load       Tokens          Owns (effective)   Host ID                                Rack
UN  192.168.220.13  9.51MiB    256             100.0%             3dba71c8-eea7-4e35-8f35-4386e7944894   rack1
UN  192.168.220.23  9.53MiB    256             100.0%             3af56f1f-7685-4b5b-b73f-efdaa371e96e   rack1
UN  192.168.220.33  9.55MiB    256             100.0%             RANDOMLY-MADE-UUID-GOES-INTHISPLACE!   rack1

A UN at the begginng of the line, refers to a node that is Up and Normal.

You can also check the logs of the cassandra server with

journalctl -u cassandra.service 

How to inspect tables and data manually

cqlsh
# from the cqlsh shell
describe keyspaces
use <keyspace>;
describe tables;
select * from <tablename> WHERE <primarykey>=<some-value> LIMIT 10;

If your local install does not have cqlsh available, you can use docker instead:

sudo docker run -it --rm cassandra:3.11 cqlsh 172.16.0.132 9042

How to rolling-restart a cassandra cluster

For maintenance you may need to restart the cluster.

On each server one by one:

check your cluster is healthy: nodetool status or nodetool -h ::FFFF:127.0.0.1 status (in newer versions)
nodetool drain && systemctl stop cassandra (to stop accepting writes and flush data to disk; then stop the process)
do any operation you need, if any
Start the cassandra daemon process: systemctl start cassandra
Wait for your cluster to be healthy again.
Do the same on the next server.