Closed magland closed 8 months ago
@luiztauffer This should fix it, but it won't take effect until it is pushed to pypi, and the app image is rebuilt with the updated version.
Attention: 20 lines
in your changes are missing coverage. Please review.
Comparison is base (
0508d2f
) 70.42% compared to head (d95532e
) 71.05%. Report is 10 commits behind head on main.
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
From slack Luiz: for some reason, when the job fails locally, the controller stops updating the output with the latest info, which includes the erro track etc
Jeremy: i know what's happening. In order to make the system more robust, I separated the console output monitor and the resource utilization monitors into separate processes. When the child job process fails (or finishes), the parent process exits, and then the container is closed before the console monitor has a chance to send final console output to dendro.
So I just need to put in a final upload of console output at the end, controlled by the parent process