hi, sorry if it is a dumb question...
I am still using lvm/iscsi storage for my vms,
on a 1TB shared iscsi/lvm storage I have many raw vm disks, but apparently many of them related to old or test machines now _deleted_
now I am thinking to remove those, to free up space there, but I wish to be shure no node is using them!!
I guess they are from old vm because of the LV name, like:
vm-112-disk-2 pve_vm_disks -wi----- 200.00g
so I guess this was created for a vm with VMID 112, and now is not "active/open" (I learned that it would show an "a" at bit 5, and "o" at bit 6 if "open")
to check this things I can log into each node and issue a lvs command, but i get a list of all disk on any node, even if that vm is not on the node.
simply, only the "owning" node will show it as "active/open" if it is so (if the VM using the LV is running there), the others will always show the LV with the two bits above as: as "--", and I don't know
I can check a couple of nodes manually, one by one, but is there any "unified" way of check this kind of things?
perhaps with distributed storage (ceph, sheepdog) is kinda easier to get this kind of cluster-wide vm-storage usage info.. ?
Marco
I am still using lvm/iscsi storage for my vms,
on a 1TB shared iscsi/lvm storage I have many raw vm disks, but apparently many of them related to old or test machines now _deleted_
now I am thinking to remove those, to free up space there, but I wish to be shure no node is using them!!
I guess they are from old vm because of the LV name, like:
vm-112-disk-2 pve_vm_disks -wi----- 200.00g
so I guess this was created for a vm with VMID 112, and now is not "active/open" (I learned that it would show an "a" at bit 5, and "o" at bit 6 if "open")
to check this things I can log into each node and issue a lvs command, but i get a list of all disk on any node, even if that vm is not on the node.
simply, only the "owning" node will show it as "active/open" if it is so (if the VM using the LV is running there), the others will always show the LV with the two bits above as: as "--", and I don't know
I can check a couple of nodes manually, one by one, but is there any "unified" way of check this kind of things?
perhaps with distributed storage (ceph, sheepdog) is kinda easier to get this kind of cluster-wide vm-storage usage info.. ?
Marco