nosql - Hardware maintenance activity on One of the Cassandra Node

Question

Need help, I have a 4 node cassandra Cluster, RF 2 and There is a Hardware maintenance activity (Total Activity time can be 30-40 mins) scheduled on one of the node . Please let me know how we can safely do this activity without impacting the live traffic.

Can I use below steps on node (where hardware maintenance will be going on)

1. nodetool -h<node IP / Hostname > drain
2. Kill Cassnadra service.
3. Once activity get completed, Then start the cassandra service.

Kindly let me know if anything else need to be done. Thanks in advance.

question from:https://stackoverflow.com/questions/65939518/hardware-maintenance-activity-on-one-of-the-cassandra-node

Answer 1

That's a good start, Dinesh. The shutdown scripts which I write look like this:

nodetool disablegossip
nodetool disablebinary
nodetool drain

The disable commands first take the node out of gossip, and then stop any native binary connections. Once those complete, I drain the node.

Once those have completed, I then stop the service.