I recently added a couple of new datastores to my VMware SAN and went about migrating some of my virtual machines to provide more space on my existing stores when I noticed something odd. I use a numeric method naming all of my datastores VMFS# starting with 1 through however many I will end up with. On VMFS1, and only VMFS1 I am seeing a number of what I can best describe as "echo" virtual machines that are located on other data stores. What I mean is that I have a virtual machine that is on VMFS3, but on VMFS1 I also see the same virtual machine, which is a single disk system. The machine was originally configured as 30GB. On VMFS3 I see that the system is provisioned as roughly the 30 GB, but it is showing as about 25GB used, and the used allocation is at its maximum. When I look at the same VM on VMFS1 I again see that 30GB is provisioned but for used I am seeing 0 of 25GB. I see this happening with about 10 virtual machines. I have another VM that on the non-VMFS1 location I have an 70GB provision and a maxed 41GB used, then again on VMFS1 I see the same thing but with a 0 in the used.
I tried experimenting with a non-critical system that was showing this duplicate. I moved the source VM which was not on VMFS1 to one of the newest datastores that I created, which it migrated successfully. The instance on VMFS1 was still there and when I attempted to move that one as well to the same datastore, it instantly ended successfully but did not move the instance. I did not try moving it to a different datastore.
Does anyone have any idea what is going on and how to consolidate these multiple instances back to a single one? This is affecting virtual machines of different sizes, usually 30-80GB, using thin provisioning. I have other VMs that are larger (150GB) also thinly provisioned that are not experiencing this issue but are also not at their "used" capacity. Thanks!
