unclear error when MAAS returns bad response during bootstrap
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| Canonical Juju |
Triaged
|
Low
|
Joseph Phillips | ||
Bug Description
While bootstrapping a controller against a local MAAS, Juju requested an instance from MAAS and then instantly said the bootstrap process was canceled. Lately one of the SQA MAAS has been a bit overloaded, and returning 400 errors during some requests. I suspect that is what happened here, however the response from Juju was unclear.
The output is as follows:
Creating Juju controller "foundations-maas" on maas_cloud/default
Looking for packaged Juju agent version 3.1.6 for amd64
No packaged binary found, preparing local Juju agent binary
Launching controller instance(s) on maas_cloud/
- k7dg4b (arch=amd64 mem=24G cores=4)
Installing Juju agent on bootstrap instance
Waiting for address
ERROR failed to bootstrap model: cancelled
Looking back in the MAAS logs for the host in the debug output, the system received no actions during the time the bootstrap was happening, which is why it leads me to think that MAAS returned an error.
The testrun can be found at:
https:/
But there were no logs collected because no instances go started to collect logs from.
| description: | updated |
| description: | updated |
| Joseph Phillips (manadart) wrote | #1 |
| Changed in juju: | |
| status: | New → Triaged |
| importance: | Undecided → Low |
| assignee: | nobody → Joseph Phillips (manadart) |
| milestone: | none → 3.1.7 |
There's not much we can do for the specific *cause* here.
It looks as though Juju successfully issues the node creation request, but the instance never becomes available as far as it knows.
The error message is due to the context timing out (cancelled). We should make that error clearer.
It isn't failing to bootstrap a *model* either it's the wholesale controller bootstrap that fails.