pvesh mem: different values between node and vm level

[darkhole]

when getting information about the memory consumption of a vm via pvesh, i get different values:

pvesh get /nodes/endor/qemu/

pvesh get /nodes/endor/qemu/206/status/current

the web interface shows the same valus like the request on vm level:


rebooting the server has no effects, the difference will stay.
the API doc tells no difference on this values. Is this an error or feature which i don't understand?
thanks in advance.

Proxmox Version:

proxmox pve-manager/8.4.17
kernel Linux 6.8.12-19-pve

 

[mfederanko]

The endpoint for a specific VM reads additional information via QMP [0]. The concrete implementation can be seen at [1], note that I linked to the source for PVE 9, but this also holds true for PVE 8.

pvesh get /nodes/endor/qemu/ --full

should yield the same information, but takes longer to execute.

[0]: https://www.qemu.org/docs/master/interop/qemu-qmp-ref.html
[1]: https://git.proxmox.com/?p=qemu-ser...b5fd36e58a58a56e7a6fe79348db652;hb=HEAD#l2489