Migrate to new team features¶

Release of wire-server-5.12 introduces a new data storage format for team features. To migrate to the new format, we have added a tool called migrate-features to our deployment bundle, which will have to be run, after upgrading to a wire-server release past 5.12. Migration has to be run only once. If you do not perform this migration after deploying to a release past 5.12, your team settings will be in an inconsistent state.

Configuration¶

The values of the chart migrate-features/values.yaml already contains the necessary configuration options.

1
2
3
4
5
6
7

job:
  name: migrate-features
  restartPolicy: Never
  backoffLimit: 4
  cassandraHost: cassandra-external # Replace with your Cassandra service name `kubectl get svc -n default`
  cassandraPort: 9042
  galleyKeyspace: galley

Most configuration values can remain as they are, but ensure the cassandraHost value is correct by running:

1

kubectl get svc -n default` # assuming the services are in the default namespace

If your cassandra installation differs from the standard wire-server-deploy setup, a hostname or an IP address of one of the cassandra nodes can be supplied here.

Run the migration job¶

When this chart is installed, it sets up a migrate-feature job in the default namespace:

1

helm install default ./charts/migrate-features

Before running the helm command, make sure the migrate-features chart is available in the charts directory, if not please download the chart or copy it from one of our latest artifact.

Troubleshoot and monitoring¶

After the helm installation check if the job is running

1

kubectl get jobs -n default

When the job is complete, you will see a migrate-features job with a COMPLETIONS status of 1/1. This may take some time. If the job does not complete, check the pod logs or Kubernetes events.

To check the pod events run:

1

kubectl describe job migrate-features -n default

To check the logs run:

1

kubectl logs job/migrate-features -n default

While the migration tool is running, team features are going to operate in read-only mode for the team that is currently being migrated. After migration, the new storage is going to be used. No other action should be required on the part of instance operators besides running the migration tool.