At least in OS, we're leaving leftover ports behind when detaching IPs and deleting instances. On our Jetstream allocation we had almost 30 ports (~1/3 of our quota) in the default network, all of which seemed to have been previously attached to old GVL instances. I'm not sure if this is a problem across the board or if it might be specific to a certain way we are deleting those instances/workers in CloudLaunch/CloudMan, but I did notice the names of almost all of them were related to GVL nodes (but that could be just a side effect of how much more we launch those vs manual instances, and I generally create a separate network and delete it for one-off instances)
At least in OS, we're leaving leftover ports behind when detaching IPs and deleting instances. On our Jetstream allocation we had almost 30 ports (~1/3 of our quota) in the default network, all of which seemed to have been previously attached to old GVL instances. I'm not sure if this is a problem across the board or if it might be specific to a certain way we are deleting those instances/workers in CloudLaunch/CloudMan, but I did notice the names of almost all of them were related to GVL nodes (but that could be just a side effect of how much more we launch those vs manual instances, and I generally create a separate network and delete it for one-off instances)