Stop template resolving if not needed #1943
Conversation
Avoiding the template ID resolving is needed, because otherwise the teardown of existing machines can fail if the named template does no longer exist. This otherwise leads to node rotation issues. Signed-off-by: Kim Fehrs <[email protected]>
kubermatic-bot
added
kind/bug
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: The full list of commands accepted by this bot can be found here.
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
kubermatic-bot
added
sig/cluster-management
|
Hi @koflanx. Thanks for your PR. I'm waiting for a kubermatic member to verify that this patch is reasonable to test. If it is, they should reply with Once the patch is verified, the new status will be reflected by the I understand the commands that are listed here. Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. |
kubermatic-bot
added
the
size/M
Upstream PRs use virtio as nic kubermatic#1942 stop template resolving kubermatic#1943 Originally implemented in - 11e2b3c - 07f3827
Upstream PRs use virtio as nic kubermatic#1942 stop template resolving kubermatic#1943 Originally implemented in - 11e2b3c - 07f3827 Stop template resolving if not needed Avoiding the template ID resolving is needed, because otherwise the teardown of existing machines can fail if the named template does not longer exist. This happened as part of the Flatcar rollout (VSO-2422) and led to node rotation issues. Closes: ANXKUBE-1361 Refactor provisioning handling of VMs Checking the returned errors alone does not match the behaviour of the API, which keeps the errors field, even if later runs of the provisioning task we're successful later. Because of that, the machine controller was not able to delete machines that were created, but not yet provisioned as nodes. By evaluating the "status" field instead of the presence of errors, we can (hopefully) better verify the provisioning status of a VM. Furthemore, the default NIC type got changed to "virtio", because of breaking API changes introduced and not yet reverted (VSSUP-16). Unfortunately, in order to work with the newer go-anxcloud release, the Go version in the Dockerfile had to be bumped as well. Tested on my playground cluster, where the provisioning of new machines is working. Closes: ANXKUBE-1326
What this PR does / why we need it:
This PR allows users to rotate their anexia nodes even if the OS template they used before has since been removed. This might happen due to security patches or similar.
Avoiding the template ID resolving is needed, because otherwise the teardown of existing machines can fail if the named template does no longer exist. This otherwise leads to node rotation issues.
Which issue(s) this PR fixes:
Not reported as an issue outside of anexia.
What type of PR is this?
/kind bug
Special notes for your reviewer:
This is only related to the cloud provider anexia.
Does this PR introduce a user-facing change? Then add your Release Note here:
Documentation: