You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Describe the issue
In #8909 we added a new k3d e2e test. When I run this locally it doesn't work, such that every time it gets past the should respond to changes even after services are restarted test, it stops running. The terminal shows an endless stream of:
Error with request: GET https://localhost/api/info 404
API check failed, trying again in 1 second
404 - "404 page not found\n"
Ubuntu 20.04 5.15.0-102-generic #112~20.04.1-Ubuntu SMP
k3d version v5.6.3
k3s version v1.28.8-k3s1 (default)
Docker version 26.1.1, build 4cf5afa
Steps to reproduce:
in terminal A, start the tests: npm run integration-all-k3d-local
wait a minute or so, then in terminal B, set the context to the local cluster: kubectl config use-context k3d-cht-e2e
in terminal B, run this watch command: watch "docker exec k3d-cht-e2e-server-0 crictl images|egrep 'TAG|medicm';echo;kubectl --namespace cht-e2e get all" Note at the top there's a list of images - there should be 6
when you tests in terminal A start to get stuck on Checking API - API check failed, trying again in 1 second - check the list of images and pod health in terminal B.
You should see less than 6 images and ImagePullBackOff errors in the pods in terminal B. This shows that the images are being removed from k3d (not longer imported)
As the images are being removed from k3d, it's assumed that imagePullPolicy: Nevermight fix it, but per my templates directory in the mrjones-imagePullPolicy branch, this didn't seem to fix it.
Describe the improvement you'd like
e2e should work locally as well as in CI
Describe alternatives you've considered @dianabarsan suggested:
If this proves to be a generalized problem, there's an option to switch to using a registry locally - and pushing new built images into the registry and instructing k3d to use it. I think it's bit more work to set up that just importing.
The text was updated successfully, but these errors were encountered:
mrjones-plip
changed the title
k3d e2e integration-all-k3d-local test fails
k3d e2e integration-all-k3d-local test fails on local dev instances
May 3, 2024
Describe the issue
In #8909 we added a new k3d e2e test. When I run this locally it doesn't work, such that every time it gets past the
should respond to changes even after services are restarted
test, it stops running. The terminal shows an endless stream of:Ubuntu 20.04 5.15.0-102-generic #112~20.04.1-Ubuntu SMP
k3d version v5.6.3
k3s version v1.28.8-k3s1 (default)
Docker version 26.1.1, build 4cf5afa
Steps to reproduce:
npm run integration-all-k3d-local
kubectl config use-context k3d-cht-e2e
watch "docker exec k3d-cht-e2e-server-0 crictl images|egrep 'TAG|medicm';echo;kubectl --namespace cht-e2e get all"
Note at the top there's a list of images - there should be 6Checking API - API check failed, trying again in 1 second
- check the list of images and pod health in terminal B.You should see less than 6 images and
ImagePullBackOff
errors in the pods in terminal B. This shows that the images are being removed from k3d (not longer imported)As the images are being removed from k3d, it's assumed that
imagePullPolicy: Never
might fix it, but per my templates directory in themrjones-imagePullPolicy
branch, this didn't seem to fix it.Describe the improvement you'd like
e2e should work locally as well as in CI
Describe alternatives you've considered
@dianabarsan suggested:
The text was updated successfully, but these errors were encountered: