Open mayrstefan opened 1 year ago
Our current workaround is the following: we have a sleep loop that checks if any process is sharing stdout with our current shell. When no more processes are left using the same file descriptor we know the updates have finished.
We use group_system_updates.py to apply updates to our systems. This iterates through all systems of a group and calls system_update.py to schedule updates for each system. After the update is scheduled group_system_updates.py exits but system_update.py is still running with file handles for stdout/stderr open. But system_update.py is no more a child process of the shell that called group_system_update.py. This is a situation that is not easy to handle. It should be one of the two options:
For the async case it would great to have an alternative way to find when all scheduled updates are finished