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
For a deploy I have, I'm looking to shut down a supervisord during the nodes upgrade because there are multiple nodes of this specific type. So I have HA and can do a "rolling upgrade" to avoid the currently upgrading server from serving requests with partially updated state.
The problem is that for a fully idempotent upgrade/install, I cannot assume that the service has even been created yet. So during the stop, it will blow up if the service doesn't exist.
Now, I can catch that exception and roll on, but the problem is that the exception is not typed, so I cannot be confident that the known exception is even what occurred as I allow it to happen while trying a stop. So...
Should there be a more specific exception type for this?
Is there a better way to try a stop where the service may not yet exist and am I just doing it wrong?
For a deploy I have, I'm looking to shut down a supervisord during the nodes upgrade because there are multiple nodes of this specific type. So I have HA and can do a "rolling upgrade" to avoid the currently upgrading server from serving requests with partially updated state.
The problem is that for a fully idempotent upgrade/install, I cannot assume that the service has even been created yet. So during the stop, it will blow up if the service doesn't exist.
Now, I can catch that exception and roll on, but the problem is that the exception is not typed, so I cannot be confident that the known exception is even what occurred as I allow it to happen while trying a stop. So...