Rolling Restart Procedure

This is a general procedure which describes how to perform a rolling restart. You can use this procedure where a restart of each node is required (changing the scylla.yaml file for example).

Note

Perform this procedure on one node at the time. Move to the next node only after validating the current node is up and running.

Procedure

  1. Run nodetool drain command (Scylla stops listening to its connections from the client and other nodes).
  2. Stop the Scylla node.

CentOS, RHEL or Ubuntu 16.04

sudo systemctl stop scylla-server

Ubuntu 14.04 or Debian

sudo service scylla-server stop

Docker (without stopping some-scylla container)

docker exec -it some-scylla supervisorctl stop scylla
  1. Update the relevant configuration file, for example, scylla.yaml the file can be found under /etc/scylla/.
  2. Start Scylla node.

CentOS, RHEL or Ubuntu 16.04

sudo systemctl start scylla-server

Ubuntu 14.04 or Debian

sudo service scylla-server start

Docker (with some-scylla container already running)

docker exec -it some-scylla supervisorctl start scylla
  1. Verify the node is up and has returned to the Scylla cluster using nodetool status.
  2. Repeat this procedure for all the relvant nodes in the cluster.