Hello everyone
We're running go binaries with supervisor on our linux boxes, some API's. So far we've been scp'ing the binaries over and restarting supervisor, that worked.
Then the apps started to get busy, and scp was failing as the file was busy, sounds sensible.
So the solution was to stop the app, then copy the file over, then start the app. Great.
apart from one thing
Another app was calling our API at the time and erred, pain.
Can anyone give me some ideas about how to do zero downtime deploys with go?
Any help, tips, links or general abuse is greatly welcomed.
Matt