LXD trust confirmation doesn't distinguish auth errors and other errors

Bug #1949949 reported by Alberto Donato
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Fix Released

Bug Description

Currently when adding a new LXD (server or cluster), the auth check keeps spinning even if authentication succeeds but a different error is returned.

The UI doesn't show what's going on.

This is what the websocket call returns in the case where the cert is not trusted:

{"type": 1, "request_id": 19, "rtype": 1, "error": "Failed talking to pod: VM Host 0: Certificate is not trusted and no password was given"}

vs the case where another error occurs (after the certificate has been trusted):

{"type": 1, "request_id": 21, "rtype": 1, "error": "Failed talking to pod: Please upgrade your LXD host to 4.16 or higher to support the following extensions: custom_block_volumes"}

The easiest way would be to show the error in both cases, so the user knows what's going on.

Alternatively we could see how to add a different error code for auth or other errors

Tags: ui Edit Tag help
Changed in maas-ui:
importance: Undecided → Unknown
Revision history for this message
Huw Wilkins (huwshimi) wrote :

Hi Alberto. I happen to have a PR in review that fixes this:


At the moment it matches the string for an auth error (which it doesn't display) but it would be great if there was an error code we could use instead to make this less error prone in the future.

Revision history for this message
Caleb Ellis (caleb-ellis) wrote :
Changed in maas:
status: New → Fix Committed
Changed in maas-ui:
status: New → Fix Released
Alberto Donato (ack)
Changed in maas:
milestone: 3.1.0 → 3.1.0-rc1
Alberto Donato (ack)
Changed in maas:
status: Fix Committed → Fix Released
To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.