Hi, Reading through recent magnum reviews on gerrit, I've noticed Spyros' comment that he hopes not many people use this option. That lead me to look into issues related to this, but the only related piece of information I could find was that this option has been observed (or proven, depending on whether we read release notes or commit message) to become a bottleneck for scaling larger clusters. Firstly, if this option is considered problematic for scaling wouldn't it make sense to somehow deprecated it, and put a warning in the documentation describing the reasoning for that? Right now it seems to be a preferred way of deploying clusters based on the documentation - it is used in the cluster template creation example here: https://docs.openstack.org/magnum/latest/user/ Secondly, if docker volume is considered problematic, does this mean that volume-based instances have the same problem in general, and image-based instances should be used instead? When does this become a problem? For clusters with 20 nodes? 100? 200? 500? -- Krzysztof Klimonda kklimonda@syntaxhighlighted.com