MAAS GUI shows incorrectly displays false failure message from prior operation

[Craig Bender]

MAAS version: 2.6.1 (7832-g17912cdc9-0ubuntu1~19.04.1)

When a node that previously failed to deploy/commission (for a correctable reason - let's say a typo in the main package repository url) gets redeployed/recommissioned, the status of that node under the machines tab of the GUI incorrectly show the last status of the prior operation. This message is display for the current operation until the "Configuring OS" status updates the message in the GUI.

This causes confusion and will result in people wrongly aborting operations that (if the error was corrected) would succeed if they just waited.

Screen caps attached.

[Craig Bender (craig-bender)]

To easily recreate this issue:

1) Edit the "Ubuntu archive" repo under Settings -> Package Repositories and change it to repo.example.com

2) Attempt to deploy a machine

3) After failure, release the machines.

4) Set "Ubuntu archive" repo back to archive.ubuntu.com/ubuntu

4) Re-attempt prior deploy operation (or commission the machine)

5) View status field of machine under 2.61 GUI

6) Note that the status message will display the "Node installation failure..." for about 5 minutes until the status message "Configuring OS" finally overwrites the false failure.

[Craig Bender (craig-bender)]

Suggest either when a machine is released, or a new operation is performed, its last status should be cleared.

$ maas admin machine read ndhfpe|jq -r '"\(.hostname)|\(.system_id)|\(.status)|\(.status_name)|\"\(.status_message)\""'

node02ob20|ndhfpe|4|Ready|"Node installation failure - 'cloudinit' running modules for final"

[Huw Wilkins (huwshimi)]

Removed maas-ui from this as the UI shows the status that is provided from the API.

[Jerzy Husakowski (jhusakowski)]

Is this still reproducible on a recent MAAS (3.3+)?