docker network connect network id on a running container does not return an error

[hickeng]

This issue is for returning an appropriate error message if attempting to connect a running container to a network.

It may be that the appropriate implementation of this is is a general one, returning an error from Commit if any guestinfo keys that are not tagged with @non-persistent are present.

Example of current behaviour:

$ docker run -dit --net=bridge alpine /bin/ash
c7abb78201ba357b7edd9c653099601a5e280239670c651b15ce8f7872b6d662
$ docker network connect other-bridge c7
$ echo $?
0
$

Notes:
For support of docker network connect against a container that is:

• running
• connected to a bridge network
• connecting to another bridge network

We must have either of:

• Switch from datastore files to namespace usage #3286 and at least migrating the use of extraconfig so that we can have non-volatile keys set when VM is powered on
• allowing disparate behaviour between vSphere versions with the volatile key bug fixed vs those without

#743 is for implementing this scenario.

For this to be viable in any of the other scenarios we need to enable hotadd for vNIC in the containerVMs.