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
When the service runned by gunicorn, the code runs in the "workers", which are "pinged" by gunicorn. If the "worker" is silent for more than the value of the timeout option, the worker will be killed and restarted.
The "worker" is currently created when the service starts and can't be pinged by gunicorn before the app.run () was called.
The problem is that all repositories are synced when the service starts (before app.run ()) because:
this sync can take a long time
it must be executed once (in one thread). So. we can't run gunicorn with workers > 1.
Both problems can be solved by starting the server by gunicorn with the preload option, but in this case we should to run
permanent sync thread after calling app.run (). Now it's run when s3model is created.
The text was updated successfully, but these errors were encountered:
When the service runned by gunicorn, the code runs in the "workers", which are "pinged" by gunicorn. If the "worker" is silent for more than the value of the
timeout
option, the worker will be killed and restarted.The "worker" is currently created when the service starts and can't be pinged by gunicorn before the
app.run ()
was called.The problem is that all repositories are synced when the service starts (before
app.run ()
) because:workers
> 1.Both problems can be solved by starting the server by gunicorn with the
preload
option, but in this case we should to runpermanent sync thread after calling
app.run ()
. Now it's run whens3model
is created.The text was updated successfully, but these errors were encountered: