-
Notifications
You must be signed in to change notification settings - Fork 90
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
feat: use an upgrade service deployment mechanism for embedded clusters #4756
Conversation
* airgap preflight reporting for upgrade service * create SubmitAppInfo func for upgrade service * use UpgradeServiceParams where possible
* ability to deploy an available update
* Airgap upgrade service --------- Co-authored-by: Craig O'Donnell <[email protected]>
* feat(ec): ability to upgrade the cluster * compile * f * f * f * f * remove test * increase upgrade timeout to 15 minutes
…lish (#4734) * wip * wip * loading state for start service minor ui changes * remove log
…hts workflow (#4752) --------- Co-authored-by: Salah Al Saleh <[email protected]>
* upgrade service integration test
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I have not adequately reviewed the TS/TSX stuff yet, but go looks ok
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Approved from front end perspective. Didn't see anything that really jumped out at me. Love to see the organizational and TypeScript improvements! Would be nice to add unit tests in the future. But no blockers from me.
Sharing the Looms that were shared with me (for posterity): |
What this PR does:
This PR changes the deployment mechanism for embedded clusters so that on upgrades, the version of KOTS / embedded cluster included in the release is used instead of the currently deployed version of KOTS / embedded cluster to fetch, render, and deploy the release.
It also changes the versioning mechanism so it is more similar to how Helm works where a new version/revision is created only when a change is actually deployed. When you deploy, you take all your inputs (manifests, config values, charts, registry settings, etc.), put them together, and deploy the new revision. There is no pre-rendering of versions that never get deployed, and you don’t have to make multiple changes to change multiple inputs.
NOTE: This new versioning / deployment mechanism is currently for embedded clusters only, and the plan is to adopt it for existing/kURL clusters later on.
Why we need it:
Main problem targeted by this change:
A key design goal for embedded cluster is that the infra and the app are coupled together, so the vendor tests and deploys their release and expects a particular version of their application to work on a particular version of embedded cluster and KOTS. During upgrades, however, the currently deployed version of KOTS pulls in and renders newly available versions. These newly available versions might specify a different embedded cluster version that includes a different KOTS version, which means the vendor might make use of new KOTS functionality. The old version of KOTS may be unable to render the newly available version, which can block the end-user from upgrading unless the vendor creates new releases that first upgrade KOTS and only later leverage the new functionality available in that version.
Other problems that are solved or improved by this approach include:
Which issue(s) this PR fixes:
Fixes SC-105486
Special notes for your reviewer:
Steps to reproduce
Does this PR introduce a user-facing change?
Does this PR require documentation?
TBD