Open aramfe opened 5 months ago
It's currently a CRITICAL issue and makes our runs completely non-functionable.
After some digging, I suppose it's because the locally generated files DO NOT get delete after a run...
Is there a way to "reset" it on the DTE agent machines and main when the nx cloud run already started in the middle of the execution?
Something like this would help here:
- run: npx nx-cloud start-ci-run
- run: yarn nx run-many -t foo, bar --args="--baz=qux"
- run: |
npx nx-cloud propagate-command-to-agents -s "yarn clean"
NX_CLOUD_DISTRIBUTED_EXECUTION=false yarn clean
- run: yarn nx run-many
Would love to hear from you
Current Behavior
When running squential commands using Nx Cloud through GitHub Actions, the command works correctly for the first attempt differentiated solely by a --args="--deployment=preproduction parameter. However, upon attempting a second deployment, especially targeting the --deployment=production environment, the process enters an infinite retry loop and can't continue folloing processes.
This behavior does not result in an error that would terminate the shell execution but instead leads to GitHub cancelling the run due to exceeding the 1-hour timeout limit. Upon inspection within the Nx Cloud interface, the problematic command appears to execute successfully (indicated by a green status), suggesting the issue arises post-execution, likely involving the Nx remote cache.
Expected Behavior
GitHub Repo
No response
Steps to Reproduce
This is how it is invoked:
Basically run a command twice in one single npx nx-cloud start-ci-run run.
Nx Report
Failure Logs
Package Manager Version
Yarn: 3.5.1 OS: Github Ubuntu Runner
Operating System
Additional Information
No response