-
Notifications
You must be signed in to change notification settings - Fork 120
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
refactor(ci): use docker instead of Konlet for GCP deployments in CI #4252
Conversation
This is a workaround for an issue related to disk partitioning, caused by a GCP service called Konlet, while mounting the cached disks to the VM and then to the container
I've confirm we'll have to use this workaround. The solutions given by GCP's support (which fixed other issues) didn't fix this one. |
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.
This is a large PR, which I'm having trouble understanding and reviewing.
It looks like the diff contains:
- changes from previous PRs in this series
- bug fixes
- functional changes
- syntax and name cleanups
I'd really like to see:
- a list of changes in this PR
- how you've tested it
- test logs that show that we haven't accidentally skipped any tests
If it's not possible to test it until it is merged to main
, can we just fix one workflow or disk first? Then we can fix the other workflows once we know the fix works.
@teor2345 I've removed everything that's not related to this PR |
After merging This is the most relevant PR to get |
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.
Alright, let's try this and see how it goes.
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.
👍
Motivation
This is a workaround for an issue related to disk partitioning, caused by a GCP service called Konlet, while mounting the cached disks to the VM and then to the container.
This has been causing most integration tests to fail when a disk with cached state is needed.
Fixes #4114
Closes #4080
Closes #4090
Depends-On: #4266
Solution
** Drawbacks **
Review
Reviewer Checklist
Follow Up Work
Follow up with GCP support ticket to confirm if there's a fix or an unknown workaround which doesn't require us to stop using Konlet.