unique vm ids across clusters?

[nktech1135]

Hello.
For those of you running multiple proxmox clusters, is there a reason to enforce unique vm ids across the entire infrastructure? or would this be an unnecessary complication?
If you do enforce unique vm ids, how do you do it?
We're working on an ansible project to automate vm/container deployment and are wondering what others use for there vm id strategy if they have one.

Thanks.

 

[oguz]

hi,

it's more of a matter of taste/preference in my opinion. there's no specific reason to enforce unique ids between clusters at the moment, since they won't interfere with each other.

We're working on an ansible project to automate vm/container deployment and are wondering what others use for there vm id strategy if they have one.

you can start with something simple. for automation purposes it helps to "code" your IDs according to a scheme in some way.

for example if you have more containers than VMs, then you could have VMIDs from 100-9999 and CTIDs with 10000-100000. if you have more VMs then you can do it the other way around.

 

[nktech1135]

it's more of a matter of taste/preference in my opinion. there's no specific reason to enforce unique ids between clusters at the moment, since they won't interfere with each other.

Actually, there is one scenario where they do interfere with each other, unless i'm reading things wrong.
Consider the following setup.
2 proxmox clusters, in separate physical locations with a site to site vpn joining the 2 locations.
A proxmox backup server with one storage array at site1.
If you try backing up both clusters to the pbs at site one the ids would conflict on the pbs server.
A workaround for this is to configure the pbs storage array as directory storage and having a separate directory per cluster.

I'm not sure if this is designed behavior or not, but i just wanted to point out one possible conflict to others before it bites them as it did me.

Thanks for all the help on this forum, it's great!