Scylla Manager 1.2 Release Announcement

The Scylla Enterprise team is pleased to announce the release of Scylla Manager 1.2, a production-ready release of Scylla Manager for Scylla Enterprise customers.

Scylla 1.2 focuses on a better control of repair tasks scope, allowing users to run repairs on a subset of Data Centers (DC), keyspaces, and tables, easier deployment, and setup, and improvements in security.

Related links:

• Download Scylla Manager – for customers only, or for limited evaluation.
• Get started with Scylla Manager 1.2
• sctool, Scylla Manager 1.2 CLI tool reference
• Upgrade Scylla Manager 1.x to 1.2
• Report issues with Scylla Manager using the Scylla Enterprise support channels

Upgrade to Scylla Manager 1.2

Read the upgrade guide carefully. In particular, you will need to redefine scheduled repairs. Please contact Scylla Support team for help in installing and upgrading Scylla Manager.

New features in Scylla Manager 1.2

Debian 8 and Ubuntu 16 packages

Scylla Manager is now available for Debian 8 and Ubuntu 16, packages will be shared in the next few days.

Improved Repair Granularity

In previous releases, one could only use Scylla Manager to run a cluster-wide repair. Starting with Scylla Manager 1.2, scheduled and ad-hoc repairs can be limited to:

• One or more DataCenters (DC), with --dc flag
• One node (host) with --host flag
• Repair with a subset of hosts with –with-hosts flag
• Repair primary or non-primary token ranges with –token-ranges [pr|npr] flag
• Keyspace (existed in Scylla Manager 1.1 with a different syntax)
• Table (existed in Scylla Manager 1.1 with a different syntax)

For example, the above elements can be used to:

• Run a local DC Repair
• Repair a specific node, for example after a node restart
• Repair only a restarted node, and all nodes with overlapping token ranges

For more on running a granularity repair see sctool 1.2 reference, and below on sctool updates

Sctool updates

• You now need to provide only one node IP when adding a new cluster. For example
  sctool cluster add --host=198.100.51.11 --name=prod-cluster
• New flags in sctool:
  • --dc, --host, --with-hosts, --token-ranges (see more granular repair above)
  • --fail-fast: Stops the repair process on the first error
• sctool allows selecting multiple keyspace and table using glob pattern matching, For example:
• Repair only US East DC:
  sctool repair --cluster mycluster --dc dc-us-east
• Repair all US DCs, but US West:
  sctool repair --cluster mycluster --dc dc-us-*, !dc-us-west
• Repair subset of Keyspaces:
  sctool repair --cluster mycluster -K mykeyspaces-*
• Repair subset of Keyspaces: sctool repair
  --cluster mycluster -K mykeyspace
• sctool task start: a new --continue flag allow resuming the task from where it last stopped, not from the start.

Security

• One of the common problems with adding a managed Scylla cluster to Scylla Manager was the task of creating a scylla-manager user and sharing the Scylla Manager public key to all Scylla node. With Scylla Manager 1.2 this task is greatly improved with the introduction of the new scyllamgr_ssh_setup script.Read more about the new script and how to manage a new cluster with Scylla Manager 1.2
• Two optional new parameters for sctool command ‘cluster add’: ssh-user and ssh-identity-file, allow you to specify a different SSH key and user per cluster
• Scylla Manager now uses HTTPS by default to access with Scylla nodes REST API. Note that Scylla REST API is still bound to a local IP only, and Scylla Manager uses SSH to connect to the node.

Monitoring

Scylla Grafana Monitoring 2.0 now includes Scylla Manager 1.2 dashboard

The following metrics have been updated in Scylla Manager 1.2

• Total run tasks metrics definition have changed the name from “status_total” to “run_total”
• In subsystem “repair” “unit” has changed the name to “task”
• In subsystem “repair” “keyspace” has been added

The following metric has been removed in Scylla Manager 1.2

• Subsystem “log” has been removed

About Scylla Manager

Scylla Manager adds centralized cluster administration and recurrent task automation to Scylla Enterprise. Scylla Manager 1.x includes automation of periodic repair. Future releases will provide rolling upgrades, recurrent backup, and more. With time, Scylla Manager will become the focal point of Scylla Enterprise cluster management, including a GUI frontend. Scylla Manager is available for all Scylla Enterprise customers. It can also be downloaded from scylladb.com for a 30-day trial.