Although there is a hibernation record, I touched the mapped hard disk (sata0) and found that it didn't stop running. It is speculated that Qunhui just stopped reading and writing the mapped hard disk, and the control right of the original physical hard disk is still in the place of PVE, so PVE should control the hibernation of the original physical hard disk:
hdparm -S 180 /dev/sdc
Parameter "180" corresponds to 15 minutes (180/ 12), and "sdc" is the drive letter of the physical hard disk, which can be viewed by the command "ls -l /dev/disk/by-id/".
This mapped disk can stop running, but it will wake up frequently. There is also a daemon "pvestatd" that often reads and writes hard disks. This process will periodically ask the status of each virtual machine, disk and container, and send it to each node to wake up the hard disk, so that it can be shut down directly:
Pvestatd stop # Deactivating this service cannot create a new virtual machine, so please use pvestatd start to start it if necessary.
You can also use lvm's metadata cache. If enabled, it will cache lvm information, and you will not read the hard disk frequently:
Modify the /etc/lvm/lvm.conf file.
Set use_lvmetad = 1.
/DSM-HDD- Hibernation -pve/