azure bootstrap failure on certain regions due to incorrect instance type being used

Bug #1761018 reported by Calvin Hartwell
32
This bug affects 6 people
Affects Status Importance Assigned to Milestone
Canonical Juju
Triaged
Low
Unassigned

Bug Description

Hi all,

Customer is having trouble bootstrapping to certain Azure DC using the default commands, it seems to work with some of their Azure AZ but not all of them, specifically northeurope/northerneurope is causing an issue and generating an error that the node type does not exist.

They managed to fix this with: juju bootstrap --constraints "instance-type=Standard_B2s"

My suggestion is to test bootstrapping on each AZ on Azure and AWS to ensure that the default bootstrap is working correctly as this is not so obvious to fix, you have to check the Azure error logs on the website. This should be automated and run per-release as to ensure it is easy to bootstrap to any of the AZ we support out of the box.

Default constraints may need to be changed for the default controller bootstrap depending on the AZ the customer picks.

John A Meinel (jameinel)
tags: added: azure-provider
removed: azure
Changed in juju:
status: New → Triaged
importance: Undecided → Medium
Revision history for this message
Nick Veitch (evilnick) wrote :
Download full text (4.4 KiB)

This still seems to be a problem on a lot of regions, and not limited to bootstrapping

Deploying to e.g. azure/ukwest works fine -

   juju bootstrap azure/ukwest cdk1
   Creating Juju controller "cdk1" on azure/ukwest
   Looking for packaged Juju agent version 2.7.4 for amd64
   Launching controller instance(s) on azure/ukwest...
    - machine-0 (arch=amd64 mem=3.5G cores=1)
   Installing Juju agent on bootstrap instance
   Fetching Juju GUI 2.15.0
   Waiting for address
   Attempting to connect to 192.168.16.4:22
   Attempting to connect to 51.11.120.37:22
   Connected to 51.11.120.37
   Running machine configuration script...
   Bootstrap agent now started
   Contacting Juju controller at 192.168.16.4 to verify accessibility...

   Bootstrap complete, controller "cdk1" is now available
   Controller machines are in the "controller" model
   Initial model "default" added

N.B. This deploys a 'Standard_D1_v2" instance type

Trying the same on the default region (centralus) yields errors:

juju bootstrap azure cdk2

Creating Juju controller "cdk2" on azure/centralus
Looking for packaged Juju agent version 2.7.4 for amd64
Launching controller instance(s) on azure/centralus...
ERROR creating instance failed, destroying: creating deployment "machine-0": resources.DeploymentsClient#CreateOrUpdate: Failure sending request: StatusCode=400 -- Original Error: Code="InvalidTemplateDeployment" Message="The template deployment failed with error: 'The resource with id: '/subscriptions/bf4b9325-b972-4693-b28d-17b3226ab202/resourceGroups/juju-controller-model-4545acce-5105-4207-8566-07258b146404/providers/Microsoft.Compute/virtualMachines/machine-0' failed validation with message: 'The requested size for resource '/subscriptions/bf4b9325-b972-4693-b28d-17b3226ab202/resourceGroups/juju-controller-model-4545acce-5105-4207-8566-07258b146404/providers/Microsoft.Compute/virtualMachines/machine-0' is currently not available in location 'centralus' zones '' for subscription 'bf4b9325-b972-4693-b28d-17b3226ab202'. Please try another size or deploy to a different location or zones. See https://aka.ms/azureskunotavailable for details.'.'."
ERROR bootstrap failed, destroying model: cannot start bootstrap instance: creating virtual machine "machine-0": creating deployment "machine-0": resources.DeploymentsClient#CreateOrUpdate: Failure sending request: StatusCode=400 -- Original Error: Code="InvalidTemplateDeployment" Message="The template deployment failed with error: 'The resource with id: '/subscriptions/bf4b9325-b972-4693-b28d-17b3226ab202/resourceGroups/juju-controller-model-4545acce-5105-4207-8566-07258b146404/providers/Microsoft.Compute/virtualMachines/machine-0' failed validation with message: 'The requested size for resource '/subscriptions/bf4b9325-b972-4693-b28d-17b3226ab202/resourceGroups/juju-controller-model-4545acce-5105-4207-8566-07258b146404/providers/Microsoft.Compute/virtualMachines/machine-0' is currently not available in location 'centralus' zones '' for subscription 'bf4b9325-b972-4693-b28d-17b3226ab202'. Please try another size or deploy to a different location or zones. See https://aka.ms/azureskunotavailable for details.'.'."
^[[A^[[BERRO...

Read more...

Revision history for this message
Ian Booth (wallyworld) wrote :

Looks like the valid SKUs are subscription specific sadly.
So unlike other clouds where there's a known list of instances types for regions, this will need to be dynamically determined each time for the given credential in use.

Changed in juju:
milestone: none → 2.8-beta1
Revision history for this message
Nick Veitch (evilnick) wrote :

Ah, thanks for that. I think I can document around the specific issues we have by recommending specifying instance types in an overlay, but yes, in general it sucks.

Ian Booth (wallyworld)
Changed in juju:
milestone: 2.8-beta1 → 2.8.1
Tim Penhey (thumper)
Changed in juju:
milestone: 2.8.1 → 2.8-next
Revision history for this message
Canonical Juju QA Bot (juju-qa-bot) wrote :

This bug has not been updated in 2 years, so we're marking it Low importance. If you believe this is incorrect, please update the importance.

Changed in juju:
importance: Medium → Low
tags: added: expirebugs-bot
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

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