Flint freezes sometimes
Feed back from ID15A.
- Flint was on the background but non-responsive
- Bliss was trying to contact it without success
Short traceback:
...
File "users/blissadm/.../plot.py", line 229, in attach_flint
...
bliss.common.greenlet_utils.Timeout: 30.0 seconds
- A manual kill was enough to fix the issue
As it would be good to have a strong robustness, could we do something to improve that?
- After a timeout
- We could kill the current flint process automatically (or with user confirmation in the shell)
- We could create a new flint process
- Then try again one time to connect
- Quality of life:
- Reducing the timeout when attaching flint could be nice