Host has 104 GB (DDR3 ECC) and about 88 GB is being used (according to the node's summary page). The running VMs themselves are configured for 56 GB total and only 40 GB of that is actually used at the moment (according to each VM's summary page). Why is there a 48 GB difference in host current usage vs total VMs' current usage?
P.S. - Only these 3 VMs are running. There are no LXC containers/instances running/set up. 2 of the VMs have GPU passthrough, if that matters (configured for 16 GB of RAM each). The VMs are stored on ZFS (4 x M.2 NVMe SSDs in RAIDZ1) and the Proxmox installation on a separate SATA SSD.
P.S. - Only these 3 VMs are running. There are no LXC containers/instances running/set up. 2 of the VMs have GPU passthrough, if that matters (configured for 16 GB of RAM each). The VMs are stored on ZFS (4 x M.2 NVMe SSDs in RAIDZ1) and the Proxmox installation on a separate SATA SSD.