1.2.1

Intended to be used with StorageOS Node versions 1.2.1 and greater.

IMPORTANT: Please upgrade your StorageOS cluster to at least version 1.2.1 before upgrading the CLI to version 1.2.1. Previous versions of StorageOS will not recognise version 1.2.1 of the CLI and will reject API requests. This version check was removed in StorageOS server 1.2.1.

New:

• #186 - Check block device contents before running mkfs
• #188 - Simplify CircleCI config

Fixed:

• #180 - Fix Red Hat builds to fall back to sha256sum if shasum not found
• #182 - Fix table-based formatting
• #185 - Fix cluster health help text
• #187 - Fix failing tests for template-based formatting

1.2.0

Intended to be used with StorageOS Node versions 1.2.0 and greater.

New:

• #164 - Adds more pleasant hostname ordering of the CLI output. Hostnames ending in numerical values will now be sorted numerically, rather that lexicographically.
• #168 - Adds maintenance mode support to the CLI. storageos maintenance enable|disable|inspect is now available.
• #171, #179 - Update and extend the functionality of the --format option. Formatting documentation is also available by running storageos [command] --format help.

Fixed:

• #174 - Cluster health is now re-enabled using a new method. The request is now handled internally on a StorageOS node container, meaning this command no-longer fails in k8s and load-balanced environments.

1.0.0

Version 1.0.0-rc3 of the CLI should be used after upgrading server nodes to 1.0.x. Most new features will require the latest version of StorageOS (currently storageos/node:1.0.0-rc5) to be running.

Deprecations:

• #163 - cluster health was removed as it did not respect the STORAGEOS_HOST environment variable and tried to connect to nodes directly. This caused problems with clusters behind firewalls and load balancers where direct access would typically be blocked. We expect this command to be fixed and re-appear in the next release. Until then, storageos node inspect --format '{{.Health}}' provides similar functionality.

New:

• #152 - node delete allows a node to be removed from management. See http://docs.storageos.com/docs/operations/decommission-nodes for details.
• #158 - Volumes are now sorted by namespace and name, not just name.
• #164 - Hostnames are now sorted better.

Fixed:

• #160 - Capacity is now reported in binary (GiB) rather than decimal (GB) format. This makes it consistent with the UI and is the industry-wide standard for reporting storage capacity.

1.0.0-rc3

Version 1.0.0-rc3 of the CLI should be used after upgrading server nodes to 1.0.x. Most new features will require the latest version of StorageOS (currently storageos/node:1.0.0-rc5) to be running.

Deprecations:

• #145 - Removed controller API object. Was replaced with node and deprecated in version 0.10.0.

This release adds:

• #139 - cluster connectivity command to verify required network connectivity between cluster nodes.
• #134 - Support for http proxies with the following environment variables: HTTP_PROXY, http_proxy, HTTPS_PROXY, https_proxy, NO_PROXY, no_proxy.
• #141 - A custom discovery service can now be specified by setting the STORAGEOS_DISCOVERY environment variable.
• #148 - licence command to view, apply and remove cluster licence.

Fixed:

• #132 - volume ls now uses overall volume health rather than just the master's health.
• #134 - CLI should default to use http:// scheme if not specified.
• #136 - TLS API endpoint support. Previously the https:// scheme was being stripped.
• #138 - Increased API timeout to 10 seconds.
• #140 - Reduced timeouts and retries for volume mounts.
• #143 - Fix parsing error when importing multiple ABAC policy files.
• #146 - Logout was failing silently when unable to determine correct host.
• #147 - Clarified usage use logout command.
• #149 - Rename Licence field from ArrayUUID to ClusterID.

1.0.0-rc2

This release contains mostly formatting fixes:

• Labels are now sorted
• Reduced width required to display output
• Send CLI version in requests and check compatibility with API version
• Add namespace flag to volume rm command
• Give error when removing non-existent node label

The CLI is also available as a container at http://hub.docker.com/r/storageos/cli

1.0.0-rc1

First release candidate for version 1.0.0. Due to some changes in the API, do not use this version with nodes that are version 0.10.0 or below.

0.10.0

Version 0.10.0 of the CLI should be used after upgrading server nodes to 0.10.0+. The CLI should mostly work with older server versions, except for node health and capacity stats.

Deprecations:

• label format for StorageOS features has changed from storageos.feature.foo=bar to storageos.com/foo=bar. Translation to the new format will occur automatically and deprecation warnings will be displayed. Translation and notices will be removed in the next minor release.
• Internal changes to rename controller API object to node.

This release adds:

• storageos logs -f streams StorageOS logs from the cluster via the CLI. This allows you to monitor the cluster without logging onto each node or implementing a central logging solution.
• Multiple API endpoints can now be specified as a comma-separated string with either -H or in the STORAGEOS_HOST environment variable. In case one endpoint is unavailable another is tried. This can be useful in environments where the API is accessed directly instead of through a load-balancer.
• Improved error messages, allowing translation of backend errors into more user-friendly messages.
• Enabled lazy init to improve mkfs performance.

Fixed:

• Panic assigning labels to an unitialized map.
• Multiple policies could not be created within a single jsonp file.
• No longer possible to create duplicate group assignments for a user.

0.9.5

This release adds:

• All unmounted volumes can now be deleted with storageos volume rm --all. Add --force to also delete mounted volumes. Use with extreme caution
• storageos install-bash-completion sets up bash shell completion.
• Fixed broken binary upload for new releases.

0.9.4

Adds storageos logs subcommand to view and update node logging configuration (requires server version 0.10). We expect this to be useful for troubleshooting where we can increase the logging on specific components to help diagnose issues.

Binaries were not released for this version.

0.9.3

This release adds:

• command shortcuts:
  • create: "c"
  • inspect" "i"
  • ls: "list"
  • update: "u"
  • rm: "remove"
  • health: "h"
  • namespace: "ns"
  • user: "u"
  • policy: "pol"
  • volume: "v", "vol"
  • node: "n"
• bash shortcut installer
  storageos install-bash-completion
• reduce timeout on version subcommand
• retry filesystem create on error
• volume inspect with no parameters returns all volumes
• fix output format to remove namespace when deleting volume by uuid
• storageos cluster health now sorts by node name
• generate file hashes for released binaries (thanks @darkowlzz)
• improved error message when volume not found