Hello, Am 20/09/2024 um 14:32 schrieb Pavel Tide: > 1) Connect via SSH to the PVE node and deploy a helper virtual machine (so > that users don't have to do it manually) > 2) Access the Proxmox VE API to perform other backup-related tasks (those > that cannot be done via SSH) > > In item #1 - the new VM deployment involved usage of root/sudo. > > In item #2 - certain tasks that are performed via API also require root/sudo. > We have managed to move those to the SSH part of the workflow, so now users > can use one non-root account to perform all necessary operations (instead of > using root or having to use two separate accounts). > > We think that in future there might be a situation where we might need a > superuser level of privileges while accessing the API, and there will be no > workaround to move the operation to the SSH part of the workflow. This will > result in forcing our joint users to use 'root' account again, which they > hate to do and also deem as an not secure practice.
Which situations/API calls would that be? It would be definitively helpful to get specifics here, as otherwise it's hard to help and also a bit hard to tell for sure if the Sys.Root privilege feature request would even help here. As that privilege would only allow current root-only API calls to be used by non-root admin accounts, but it would not allow the account to gain root access on the system just by having that privilege. In general, I think it would be better to do less, not more, stuff manually in the long term and rather check out the in-development external backup provider API [0], as that would allow easier and safer access to VM and CT data while integrating better with the existing PVE stack, ideally reducing the potential for fallout on either site. [0]: https://lore.proxmox.com/pve-devel/20240813132829.117460-1-f.eb...@proxmox.com/ - Thomas _______________________________________________ pve-devel mailing list pve-devel@lists.proxmox.com https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
