I was able to reproduce the problem on a mutlinode devstack started from master. Marking this bug as Confirmed.
As you reported that the same problem can be observed in Queens I don't mark it as an RC2 potential for Victoria as it is not a Victoria regression.
I was able to reproduce the problem on a mutlinode devstack started from master. Marking this bug as Confirmed.
As you reported that the same problem can be observed in Queens I don't mark it as an RC2 potential for Victoria as it is not a Victoria regression.