[HDDS-4989] Decommission CLI should return details of nodes which fail

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.2.0
• Component/s: SCM, SCM Client
• Labels: pull-request-available
• Target Version/s: 1.2.0
• Assignee: Stephen O'Donnell
• Reporter: Stephen O'Donnell
• Created: 16/Mar/21 20:39
• Updated: 08/Apr/21 15:03
• Resolved: 08/Apr/21 15:03

Description

With the current decommission / recommission / maintenance mode commands, you can pass a list of hosts to perform the operation on. If any of these hosts fail to enter the decommission / maintenance workflow, the command gives no feedback about the error. Some of the hosts can silently fail and the only way to know is to inspect the SCM log.

The most common way a host can fail, is if a node which is undergoing maintenance is instructed to go to decommission and vice versa as this is a transition which is not allowed.

This change will allow any failed nodes to feed back to the client. If the client detects that any of the nodes have failed, details will be written to stderr and the command exit code will be non-zero.

Note that even though the exit code is non-zero, the command may have partially worked.

Also note that the errors which are fed back are only around transitioning the node into the admin workflow - it is still possible for it to fail later for other reasons which will not be fed back to the client. This is because the client does not wait for the process to complete, but exits after confirmation the command has been processed by scm.

Issue Links

• links to: GitHub Pull Request #2057

Sub-Tasks

• Introduce a CLI command to show the datanodes under decommissioning: Closed