Wire-Server `5.26.0` release¶

Federated environments: skip this version.

The 5.26 charts have unresolved issues in federated mode. The original release PR was closed without merging. Federated deploys should go straight from 5.25 to 5.27.

Non-federated deploys can apply 5.26 if they want, but there are no breaking changes between 5.25 and 5.27, so skipping is also fine.

Reference based on these build.json charts, and the release changelog.

Artifact: wire-server-deploy-static-6ccd1d01db71f30efa1164cf0f9fb6c1f6f5bf64.tgz

Heads up¶

Coming from 5.25.0. For deploys at any earlier version, do the 5.25 page first.

Notes from the previous release¶

The background-worker chart default for postgresMigration.conversation is corrected at this release. It defaults to cassandra again, matching the galley default. Earlier versions of the chart defaulted to postgresql, which is exactly the bug that broke conversations on installs that hadn't migrated.

If postgresMigration.conversation was already set explicitly in values/wire-server/values.yaml (which 5.25 told operators to do), no action.

If the chart default at 5.25 was being relied on, and the migration was already done, the value has to be set explicitly to postgresql now, otherwise the new cassandra default flips the worker back to reading from Cassandra:

background-worker:
  config:
    postgresMigration:
      conversation: postgresql

What must change¶

1. Make sure the Kubernetes cluster is at `1.27` or newer¶

Anything below 1.27 is no longer supported. On older clusters, Kubernetes itself has to be upgraded before touching wire-server.

2. Re-fill the Elasticsearch index from Cassandra (after the upgrade)¶

User search now returns user type info (regular, app, legacy bot). The new fields don't show up in search results until the Elasticsearch index is re-filled from Cassandra. Do this after the wire-server upgrade.

For deploys that call brig-index directly instead of letting the chart run it, the tool needs PostgreSQL access now, on top of its existing settings. The invocation has to be updated accordingly.

See Refill ES documents from Cassandra for the actual procedure.

3. Run the wire-server helm upgrade¶

d helm upgrade --install wire-server ./charts/wire-server --timeout=15m0s \
  --values ./values/wire-server/values.yaml \
  --values ./values/wire-server/secrets.yaml

Post-upgrade: migrate conversation codes to PostgreSQL (optional)¶

Back up before starting. Take a backup of the Cassandra galley keyspace and of the target PostgreSQL database before running any of the steps below. The migration writes to PostgreSQL from step 1 onwards, and rolling back without a backup is not straightforward.

Same shape as the conversation data migration from 5.24. Only makes sense when conversation data has already been migrated, otherwise leave it.

Step 1, in values/wire-server/values.yaml:

galley:
  config:
    postgresMigration:
      conversationCodes: migration-to-postgresql
background-worker:
  config:
    postgresMigration:
      conversationCodes: migration-to-postgresql
    migrateConversationCodes: false

Run the helm upgrade.

Step 2, set migrateConversationCodes: true on background-worker and run the helm upgrade again. Wait for the wire_conv_codes_migration_finished metric to hit 1.0.

Step 3, switch reads to PostgreSQL:

galley:
  config:
    postgresMigration:
      conversationCodes: postgresql
background-worker:
  config:
    postgresMigration:
      conversationCodes: postgresql
    migrateConversationCodes: false

Run the helm upgrade.

Disk space note¶

Each upgrade in this series re-runs setup-offline-sources, which copies the new release's binaries, container images, and debs into /opt/assets on the assethost. After a few versions, the assethost runs out of space and the playbook fails with no space left on device.

When that happens, SSH into the assethost (not the adminhost) and clear it:

sudo rm -rvf /opt/assets

Then re-run setup-offline-sources from the adminhost.

Wire-Server 5.26.0 release¶