4

For some unknown reason, the manager host which gets the command to launch a stack, tries to pile up all the containers on itself; other worker/manager hosts stay unemploed.

Unfortunately there are no log entries for that.

Has anybody observed this behaviour already with Docker CE as well?

Ta Mu
  • 6,772
  • 5
  • 39
  • 82

1 Answers1

1

I would check that all nodes are active and ready with command

docker node ls

With 

docker swarm leave

and

docker swarm join \
--token <SWMTKN token> \
<manager private IP>

you may rejoin malfunctioning node.

mico
  • 525
  • 1
  • 5
  • 12