My goal is to perform rolling update of the service deployed in docker swarm. Several service tasks are in starting state by design. So they are in starting state before the update and should be in this state after. There are service tasks in running state too.
I’m using docker service update with arguments:
docker service update --image new_image:tag
The first old task in starting state is shutdown and new stays in starting state forever. So the update doesn’t continue.
Using
docker service ps service_name
clearly shows it and running
docker service inspect service_name
shows that the update is still in progress.
Is it possible to instruct docker service update command to continue rolling update of a service after timeout?