Closed jaredtbates closed 1 year ago
Hi @jaredtbates
As a quick patch, you should be able to resolve this by running npx nx run XXX:build --skip-nx-cache
from the same branch that your pipeline is running against. This will force Nx Cloud to overwrite any previous entries and solve this problem short term.
For additional information, are you using the nrwl/ci
workflows, or something more custom to your repo? Which OS are your action runners using?
Thanks @StalkAltan, I'm trying that now. We're using a slight variation of the nrwl/ci
workflows, but I already compared them and didn't see any difference that would cause this issue. I can share those publicly, since there's no sensitive info in them. Let me know if you see anything that I may have missed. We're using the default action runner OS, which I believe is ubuntu-latest
. Thank you!
It appears the issue may have been because of implicitDependencies
. For my project XXX
I also had a project XXX-e2e
. The 404s seemed to occur regularly on the XXX-e2e
e2e
target. Digging into the verbose logs, I could see that the failing build
target was under that e2e
target. I looked at my project.json
and found that the app XXX
was listed as an implicit dependency of XXX-e2e
. This was creating unnecessary builds of my project (using the XXX:build
target), because the e2e
target would build the base project as part of the XXX:serve
that happens during the e2e
execution.
I see this came from the cypress generator here: https://github.com/nrwl/nx/blob/d5c93bfa39c798bc5f594aa85e37a8b0cbe0b44e/packages/cypress/src/generators/cypress-project/cypress-project.ts#L115
We're currently facing this issue as well. Do you have any additional information on how to track down and avoid these implicit dependencies?
The implicitDependencies are not the problem. Please update your @nrwl/nx-cloud package to the latest version.
Please collect all your CI logs from this occurrence and send us an email at cloud-support@nrwl.io
You will also want to enable the NX_VERBOSE_LOGGING=true environment variable on all of your pipelines.
We are almost done with a rework that will prevent this class of problems entirely.
Going to close this issue as this is a workspace-dependent problem.
For future people that find this issue, rather than comment here, please follow the instructions above and we will be in touch.
@StalkAltan Hi. The issue is not gone. Something went wrong with the caching hashes, and NX now tries to retrieve a cache that doesn't exist. Skipping the cache didn't force-override the cache in the cloud. We need something that would let us clear the cache in the cloud as well.
This issue has been closed for more than 30 days. If this issue is still occuring, please open a new issue with more recent context.
Current Behavior
We are receiving 404s when using NX Cloud Distributed Tax Execution in our GitHub Actions. This just started a few days ago - I do not think it's related to our project setup, it seems to be a server-side issue.
Expected Behavior
No 404 and better handling of error, continuing without cache.
Github Repo
No response
Steps to Reproduce
This started out of the blue and is occurring randomly on our GitHub Actions workflows. I'm not sure exactly how to reproduce it.
Nx Report
Failure Logs
Additional Information
No response