We recommend always using “tagged” versions of StorageOS rather than “latest”, and to perform upgrades only after reading the release notes.
The latest tagged release is
installation instructions please see our
The latest CLI release is
v2.1.0, available from
To upgrade from version 1.x to 2.x, please contact support for assistance.
v2.1.0 - Released 2020-06-26
- csi: Volume expansion now supported in offline mode. To expand a volume,
stop any workloads accessing the volume, then edit the PVC to increase the
capacity. For more information, see our Volume
Resize operations page and the
CSI Volume Expansionpage.
- api: Volume configuration including replica count can now be updated while the volume is in use. Other updateable fields include labels and description.
- failover: Before determining that a node is offline and performing recovery operations, the IO path is also verified. This provides more robust failure detection and ensures that nodes that are still responding to IO do not get replaced. This IO path verification is in addition to the gossip-based failure detection.
- operator: Default tolerations are now set for the StorageOS node container. This helps ensure that the StorageOS node container does not get evicted when the node is running low on resources.
- api: Added checks to prevent deletion of a node with active volumes, or if it is the master of at least one volume. This helps prevent orphaned volumes.
- cli: Add an
--offline-deleteflag to allow removal of volumes whose master and replica nodes are offline. This allows cleanup of orphaned volumes.
- ui: Add an offline volume delete option.
- ui: Volumes can now be detached from the UI.
- cli: Labels are no longer truncated.
- api: When a new node is added to the cluster, its capacity is available to use immediately.
- ui: Favicon was missing.
- ui: Duplicate volumes could be shown on the node details page.
- operator: During uninstall a ClusterRoleBinding was not removed.
v2.0.0 - Released 2020-05-05
- operator: StorageOS containers now run in the
kube-systemnamespace by default to allow the
system-node-criticalpriority class to be set. This instructs Kubernetes to start StorageOS before application Pods, and to evict StorageOS only after application Pods have finished. This setting was previously recommended in documentation; it is now the default.
- operator: StorageOS CSI helper containers now run as privileged. This ensures that the CSI endpoint can be seen on systems with SELinux enabled.
- ui: replication progress for new or re-joining replicas is now displayed.
- ui: show warning for unlicensed clusters.
- cli: new commands:
- licence management
- get policy
- create namespace
- create policy
- describe user
- describe namespace
- describe policy
- delete user
- delete namespace
- delete policy
- licence: removed the default licence expiry date added for
- dataplane: improved retry behaviour for network IO.
- cli: “get volumes” for all namespaces should be done in parallel.
- cli: help text document config file
- ui: link node name and get to node details on the volume details page.
- ui: node details add available capacity spinner.
- ui: node list remove capacity values / address port.
- ui: node list show master/replica counts.
- ui: node list remove edit action.
- ui: format entity labels.
- ui: node details link volumes.
- ui: align buttons for licences.
- ui: k8s warning in “create volume” modal.
- ui: node list remove “API” from “API Address”
- ui: add some details about the Licence on the licence page.
- api: include valid for duration in login response.
- licence: restrict nodes which are unregistered after 24 hours.
- scheduler: return error for namespace/volume not found
- dataplane: start gRPC threads separately from rest of the supervisor.
- ui: centre licence types.
- ui: capacity in ui is per namespace.
- cli: fail gracefully if missing some output details (i.e. no node exists for id).
v2.0.0-rc.1 - Released 2020-03-31
Initial release of version 2.x. See StorageOS v2.0 Release Blog for details.