Multi Datacenter Management
- Thread starter blinds_eye
- Start date
There is some progress which can already used via CLI.... but I am still at early Beta around 2024 expectation.......
yeah - the remote migration part got applied as experimental preview (
qm remote-mgirate ... / pct remote-migrate), and we also did a lot of preparation for part re-use in other areas.That's great news all round.
Myself, I have very little interest in migration between 'datacenters' - but even just being able to view the stats between different clusters (or single node clusters too) in different locations.
My hobby stuff has had machines donated in two different datacentres around the world - and it'd be awesome to use my home test bench in the same UI as everything else.
On a professional level, it seems like a killer feature to be able to at least administer different physical locations that would normally operate independently - from the one UI. It's easy enough to get IP connectivity between the different physical locations and ACLs made to work - but I feel that this kind of featureset would be a game changer for proxmox
Myself, I have very little interest in migration between 'datacenters' - but even just being able to view the stats between different clusters (or single node clusters too) in different locations.
My hobby stuff has had machines donated in two different datacentres around the world - and it'd be awesome to use my home test bench in the same UI as everything else.
On a professional level, it seems like a killer feature to be able to at least administer different physical locations that would normally operate independently - from the one UI. It's easy enough to get IP connectivity between the different physical locations and ACLs made to work - but I feel that this kind of featureset would be a game changer for proxmox
You can do that already now by setting up metrics server and, e.g., slap on a grafana setup for viewing them.
https://pve.proxmox.com/pve-docs/chapter-sysadmin.html#external_metric_server
We use that quite successfully for monitoring our infra and CDN, which are hosts that are quite spread over the whole world.
Definitively, that's why we're working on that.
hahaha yeah - I do that too - but I use node_exporter on the proxmox setup and prometheus / grafana as well. It's the workaround I've come to use
another huge +1 for this feature. Would love an update/wonder if there is a better place to track this than this thread? (I.e. is there a feature request system I can watch this feature move through/see comments on?)
If something will be released then we'll definitively also update this thread, so you'll get a notification from the forum then too.
May I ask if something like bulk editing or syncing of security groups, IP sets and aliases is a planned feature?
Because my homelab got 4 unclustered nodes (cluster doesn't make much sense as I can't use migration anyway because of ZFS encryption and only one of those nodes is running 24/7), all nodes need the same IP sets, aliases, security groups so I can move the guests between nodes using PBS backup+restore and it is really a pain to edit these 4 times for everything I do.
If that is not a planned feature I would need to script something myself to sync the datacenter.fw between nodes.
Because my homelab got 4 unclustered nodes (cluster doesn't make much sense as I can't use migration anyway because of ZFS encryption and only one of those nodes is running 24/7), all nodes need the same IP sets, aliases, security groups so I can move the guests between nodes using PBS backup+restore and it is really a pain to edit these 4 times for everything I do.
If that is not a planned feature I would need to script something myself to sync the datacenter.fw between nodes.
Last edited:
Probably nothing for the initial version, but surely something that should be evaluated later on.
sshfs + systemd unit triggered copy or the like ;-P
Last edited:
Thanks so much for this update. I suspect you guys already know this, but I think there is a ton of new business here for you if you had this feature because I think there are a lot of people in my situation. My situation is I have been using proxmox at home for some time (keeping all your features in the community edition and the community edition mostly rock solid is basically the best advertising you could be doing IMO). At work I have basically been looking for an excuse to give you guys money because I am so tired of dealing with vmware (their solution can be highly annoying for small shops like mine that doesn't really need shared storage or a big cluster, it just needs a few simple vms for handling stuff like the microsoft ecosystem stuff, various security tools like tenable scanner, a few custom apps, etc). If I had a single pane to manage multiple clusters I could roll you guys out to branch offices as we open them and demonstrate this works, the backup server works, etc. which would give me the buy in to swap the main cluster at the head office over at a time that makes sense, likely at a major hardware upgrade. Without a way to easily monitor, apply global settings, and move/clone vms between 5-10 branch office nodes plus the main cluster at the office I can't really push this forward. It's basically the missing piece for me to start sending you cash now that the backup server is operational. EDIT: the backup server comment is only partially true. It would be amazing if you guys had a windows agent like how Datto's backup solution works, but that's not a deal breaker for me.
Last edited:
Still what I said:
.... but this will not happen before 2025.... maybe an early beta by 2024....
some first steps can be made by CLI....
Im sure Proxmox is in active development but good things need time ;-) No one wants a buggy feature that might hurt your infastructure.
In the meantime theres some external tool, but its Windows only :-( https://cluster-manager.fr/
In the meantime theres some external tool, but its Windows only :-( https://cluster-manager.fr/
yes, there is progress, but it's slow and steady e.g., Wolfgang revamped and extracted the Rust-based auth code and other backend parts so that we can re-use them, there have been some smaller improvements for remote migration, etc.pp.. like @jsterr said - we'd rather have something that works well but takes a bit to implement.
e.g., Wolfgang revamped and extracted the Rust-based auth code and other backend parts so that we can re-use them, there have been some smaller improvements for remote migration, etc.pp.. like @jsterr said - we'd rather have something that works well but takes a bit to implement.
That's not an update as of November 2023, that's from ancient times as far as development is concerned.