The agent sometimes calls into libnetwork code that in turn calls
(*Cluster).IsAgent and (*Cluster).IsManager. These can cause the
node shutdown process to time out, since they wait for a lock that is
held by Cleanup.
It turns out c.mu doesn't need to be held while calling Stop. Holding
controlMutex is sufficient. Also, (*nodeRunner).Stop must release
nodeRunner's mu during the node shutdown process, otherwise the same
call into Cluster would be blocked on this lock instead.
Signed-off-by: Aaron Lehmann <aaron.lehmann@docker.com>
Currently these fields are included in the response JSON with zero
values. It's better not to include them if the information is
unavailable (for example, on a worker node).
This turns Cluster into a pointer so that it can be left out.
Signed-off-by: Aaron Lehmann <aaron.lehmann@docker.com>
… in order to remove duplication.
Each time we update a cluster object, we do some common
operations (lock, verify it's on a manager, get the request context,
and the update). This introduce a method and refactor few
update/remove method that allows to duplicate less code.
Signed-off-by: Vincent Demeester <vincent@sbr.pm>
