Bandwidth limit not working?

[altano]

I'm trying to limit my entire Proxmox host (v8.4.5) to 500 Mbit/s (~59.6 MiB/s). I set the datacenter bwlimit to 59 MiB/s:

# cat /etc/pve/datacenter.cfg

bwlimit: default=60416

I then tried in a container on my SDN vnet to run a speedtest and it went (almost) full link speed (~900 Mbit/s up/down):

CT is using Debian 12.7-1 template. Speedtest was run using:

# sudo apt install speedtest-cli
# speedtest-cli --secure
Retrieving speedtest.net configuration...
Testing from ColoCrossing (107.174.80.244)...
Retrieving speedtest.net server list...
Selecting best server based on ping...
Hosted by WebNX Inc. (Los Angeles, CA) [0.73 km]: 2.231 ms
Testing download speed................................................................................
Download: 939.58 Mbit/s
Testing upload speed......................................................................................................
Upload: 925.53 Mbit/s

I tried in another container which was just on my vmbr0 bridge and that had the same issue, so it's not the SDN.

tc qdisc output:

# tc qdisc
qdisc noqueue 0: dev lo root refcnt 2
qdisc mq 0: dev eno1 root
qdisc pfifo_fast 0: dev eno1 parent :4 bands 3 priomap 1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1
qdisc pfifo_fast 0: dev eno1 parent :3 bands 3 priomap 1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1
qdisc pfifo_fast 0: dev eno1 parent :2 bands 3 priomap 1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1
qdisc pfifo_fast 0: dev eno1 parent :1 bands 3 priomap 1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1
qdisc pfifo_fast 0: dev tailscale0 root refcnt 2 bands 3 priomap 1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1
qdisc noqueue 0: dev vmbr0 root refcnt 2
qdisc noqueue 0: dev vnet0 root refcnt 2
qdisc pfifo_fast 0: dev tap101i0 root refcnt 2 bands 3 priomap 1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1
qdisc noqueue 0: dev veth100i0 root refcnt 2
qdisc noqueue 0: dev fwbr100i0 root refcnt 2
qdisc noqueue 0: dev fwpr100p0 root refcnt 2
qdisc noqueue 0: dev fwln100i0 root refcnt 2

 

[Hannes Laimer]

Hey,

[1] are datacenter options. So this if for the things listed there. This is not for the BW of a NIC attached to a CT or VM. For that you have to set the limit on the NIC itself(VM > Hardware > Network Device, or CT > Network under "Advanced" on both).

[1] https://pve.proxmox.com/pve-docs/pve-admin-guide.html#_options

 

[altano]

What does “datacenter options” mean? I thought it was overall bandwidth of the host. Is it something else?

If I set it for individual VM/CT, it won’t be an aggregate limit. Is there any way to create an aggregate limit for the whole host?

 

[Hannes Laimer]

Like in the UI, these are limits for those specific things. Backup/Restore, Migrations, Cloning and Disk Move. With default, like the placeholders indicate, you set all (of these) to the same.

Well, no. What would that even mean? All NICs present? Also virtual bridges? What about traffic between VMs on the same host? These limits are configured per (virtual) NIC. If you have a use-case for for having a "global" thing, you can create a feature request over at [1], and if enough people deem this useful, chances are we'll implement something like this in the future. But also make sure to be specific, things like scope and what exactly should be covered and what not.

[1] https://bugzilla.proxmox.com/