-
Notifications
You must be signed in to change notification settings - Fork 349
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
Remove nbctl daemon and fix upgrade tests #2707
Conversation
ae32de6
to
c9b5a23
Compare
c9b5a23
to
c0dbb40
Compare
c0dbb40
to
2a10828
Compare
I expect upgrade tests to fail here since the ovnkube.sh script is not upgraded in these tests, so we still try to check that the ovn-nbctl daemon is running in
Not really sure how/if I should handle this upgrade scenario quite yet |
e2a8baa
to
e8acf58
Compare
Had to fix some things around upgrade, specifically,
|
e8acf58
to
688fbce
Compare
502cf29
to
ba216a7
Compare
/retest |
3 similar comments
/retest |
/retest |
/retest |
62c5403
to
c52894c
Compare
@jcaamano If you get the chance you could also take a look here, pretty much the wrap up to libovsdb conversion |
/retest |
/retest failed |
/retest-failed |
c369060
to
142135f
Compare
/retest-failed |
/retest |
.github/workflows/test.yml
Outdated
if [ -f ${CI_IMAGE_CACHE}${CI_IMAGE_MASTER_TAR}.gz ]; then | ||
cp ${CI_IMAGE_CACHE}/${CI_IMAGE_MASTER_TAR}.gz ${CI_IMAGE_MASTER_TAR}.gz | ||
gunzip ${CI_IMAGE_MASTER_TAR}.gz | ||
echo "::set-output name=LOAD_MASTER_IMAGE_TO_CACHE::false" |
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.
nit:
It would have been much easier for me to understand if we had an MASTER_IMAGE_RESTORED_FROM_CACHE
instead of LOAD_MASTER_IMAGE_TO_CACHE
MASTER_IMAGE_RESTORED_FROM_CACHE
would be set to false by default and to true within this if clause.
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.
Ack changed the variables to MASTER_IMAGE_RESTORED_FROM_GHCR
and MASTER_IMAGE_RESTORED_FROM_CACHE
.github/workflows/test.yml
Outdated
|
||
- name: Cache master image | ||
if: steps.is_master_image_build_needed.outputs.MASTER_IMAGE_RESTORED != 'true' && success() | ||
continue-on-error: true | ||
if: (steps.is_master_image_build_needed.outputs.LOAD_MASTER_IMAGE_TO_CACHE == 'true' || steps.is_master_image_build_needed.outputs.MASTER_IMAGE_RESTORED != 'true') && success() |
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.
nit continued:
and then here we would check for MASTER_IMAGE_RESTORED==true && MASTER_IMAGE_RESTORED_FROM_CACHE==false
It just all reads more natural this way.
.github/workflows/test.yml
Outdated
- name: Check out code into the Go module directory | ||
uses: actions/checkout@v2 | ||
|
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 could not come up with a reason on why we need this checkout
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.
+1 I don't believe we do
b1a73b3
to
e2d9865
Compare
/retest |
2 similar comments
/retest |
/retest |
e2d9865
to
c9149a0
Compare
/retest-failed |
1 similar comment
/retest-failed |
/retest |
/retest-failed |
@jcaamano CI has been going a bit bonkers and is definitely not retesting as desired....Let me try and manually rebase |
Remove the ability/config to run nbctl in daemon mode in ovn-kubernetes. This is due to the fact that 90% of nbctl calls have been converted to using libovsdb clients in the ovnkube-master process rendering the daemon obsolete. Also remove the `skipped_nbctl_daemon_total` metric. Signed-off-by: astoycos <[email protected]>
Fix the GH actions workflow to ensure we use the MASTER kind scripts and images when bringing up the pre-upgrade kind cluster And checkout the PR's code to ensure we build/deploy updated manifests as well as images. Upgrade the script to properly upgrade the ovnkube resources, these include the - ovnkube-master deployment - ovnkube-db and ovnkube-node daemonsets rather than simply upgrading the ovn-image used by the resources Run go-mod tidy to clean-up some stale artifacts in go-controller/go.sum Signed-off-by: astoycos <[email protected]>
Now that ovn-k images are automatically built for master and pushed to https://github.com/orgs/ovn-org/packages We don't need to rebuild the master branch based images everytime CI runs This PR edits the `build-master` workflow to 1. Look in the cache for the image 2. If not there try and pull from ghcr.io 3. If pull fails build the image Signed-off-by: astoycos <[email protected]>
c9149a0
to
ac7ea72
Compare
/retest-failed |
Yeah, it looks like the latest workflow run is not registered and when we retest-failed it triggers an old run. |
Also the junit report jobs seems to be crossposting |
Remove the ability/config to run nbctl in daemon mode
in Ovn-Kubernetes. This is due to the fact that 90%
of nbctl calls have been converted to using libovsdb clients
in the ovnkube-master process rendering the daemon obsolete.
Would like this to merge after #2697