The current system experiences abrupt crashes in the pod, specifically when the submission.csv file is being created within the submission container. The creation time of submission.csv varies based on participant or host team submissions, leading to synchronization issues with the monitoring container.
As a temporary solution, hosts are manually adding a sleep command after submission file creation to allow the monitoring container sufficient time to pick up the file. However, this approach is considered hacky and not a sustainable solution.
Expected Behavior:
The system should handle the creation of the submission.csv file more robustly, ensuring that the monitoring container doesn't encounter issues or crashes when the file becomes available. A more elegant and automated solution is sought to replace the current workaround.
Current Behavior:
Abrupt crashes in the pod occur when the submission.csv file is being created. The monitoring container is killed, likely due to the file not being immediately available for processing.
We need to investigate and implement a cleaner solution to handle the timing discrepancies between the submission file creation and the monitoring container's activity. This may involve enhancing the synchronization mechanism or introducing a more robust way for the monitoring container to detect the availability of submission.csv.
The current system experiences abrupt crashes in the pod, specifically when the submission.csv file is being created within the submission container. The creation time of submission.csv varies based on participant or host team submissions, leading to synchronization issues with the monitoring container.
As a temporary solution, hosts are manually adding a sleep command after submission file creation to allow the monitoring container sufficient time to pick up the file. However, this approach is considered hacky and not a sustainable solution.
Expected Behavior: The system should handle the creation of the submission.csv file more robustly, ensuring that the monitoring container doesn't encounter issues or crashes when the file becomes available. A more elegant and automated solution is sought to replace the current workaround.
Current Behavior: Abrupt crashes in the pod occur when the submission.csv file is being created. The monitoring container is killed, likely due to the file not being immediately available for processing.
We need to investigate and implement a cleaner solution to handle the timing discrepancies between the submission file creation and the monitoring container's activity. This may involve enhancing the synchronization mechanism or introducing a more robust way for the monitoring container to detect the availability of submission.csv.