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
We are observing an issue in our deployment pipeline, where iron register logs an error on the console, but still seems to return a zero status code, which leads the CI agent to believe the command was successful.
There are the logs we observe (obfuscated a few bits):
iron register --name ••••-pdf-report-systest $docker_image
[command]/bin/bash ./.tempexecution.sh
-----> Configuring client
Project '0Fna••••O93••••2' with id='••••94e••••ca500••••5244'
-----> Registering worker '••••-pdf-report-systest'
400 Bad Request: Image 'docker.••••.••••.io/s•••e-••••_••••-dev/••••-pdf-report:0.0.1.1••33' does not exist or authentication failed. If this image is private or on a non-docker hub registry, make sure `iron docker login` worked and the registry is available. Last error: Get https://docker.••••.••••.io/v2/s•••e-••••_••••-dev/••••-pdf-report/manifests/0.0.1.1••33: http: non-successful response (status=401 body="{\"errors\":[{\"code\":\"UNAUTHORIZED\",\"message\":\"authentication required\",\"detail\":[{\"Type\":\"repository\",\"Class\":\"\",\"Name\":\"s•••e-••••_••••-dev/••••-pdf-report\",\"Action\":\"pull\"}]}]}\n")
Upon looking into the source code, and with my limited understanding of go, this line seems to be the culprit:
Our desired behavior would be that if the command did not execute successfully, we would obtain a non-zero return code, so the agent could detect there was an issue an fail the pipeline (giving us the power to choose whether to continue execution on errors or not).
Let me know if there is an easy workaround for this issue. I guess given the current implementation I will have to workaround it by checking the content of stdout with the success expression.
The text was updated successfully, but these errors were encountered:
Hello,
We are observing an issue in our deployment pipeline, where
iron register
logs an error on the console, but still seems to return a zero status code, which leads the CI agent to believe the command was successful.There are the logs we observe (obfuscated a few bits):
Upon looking into the source code, and with my limited understanding of go, this line seems to be the culprit:
ironcli/commands.go
Line 801 in e32f43a
Our desired behavior would be that if the command did not execute successfully, we would obtain a non-zero return code, so the agent could detect there was an issue an fail the pipeline (giving us the power to choose whether to continue execution on errors or not).
Let me know if there is an easy workaround for this issue. I guess given the current implementation I will have to workaround it by checking the content of stdout with the success expression.
The text was updated successfully, but these errors were encountered: