Proxmox Datacenter Manager - First Alpha Release

Really interesting project.
If I may suggest which features you should focus on, these are the ones I would most like to have:
  1. Improve VM and LXC migration by overcoming current limitations.
  2. Improve Remotes cluster management.
    In particular, if the pve node that was connected to the pdm is the one that is no longer active, the entire cluster will not be visible in pdm.
  3. Integration of Proxmox Backup Server and Proxmox Mail Gateway.
Keep up the good work!
 
pixelpeter said:
I don't understand the answer correctly.
When adding a node there are only two fields. The IP + port and the fingerprint.
We can't go any further here.
No matter which fingerprint I take from the three possible certificates of the node.
Click to expand...
You can try adding the CA like the above poster suggested, but that didn't work for me. Here's what I did:
- In your PVE Manager, go to Administration --> Syslog. You'll see two lines that have fingerprints. One will say "bad fingerprint" and the next will say "expected fingerprint". The expected fingerprint is likely from your CA. You can confirm this using the openssl command by doing:
Bash: 
openssl x509 -noout -fingerprint -sha256 -inform pem -in YOUR_CA.pem
this is assuming your CA is in PEM format.

If you confirm that the "expected fingerprint" is the one from your CA, then use that fingerprint in the step where you're running into issues, and again in the final step where you're asked to add nodes.

I'm pretty convinced this is not the right solution, but it's the one that works. Hoping a Staff member can chime in on this. My best guess is that the Datacentre is configured to look for the wrong fingerprint, so either we should be asked to only upload the certificate (vs the full chain) or make sure that the "expected fingerprint" is from the actual certificate, rather than the CA.
 
  • Like
Reactions: Paolo Saudin and RocketSam
I will admit i did not go through every page here yet, but if I need to re-install PDM, is there a file or somewhat to bring back all my nodes without having to re-add them 1x1? We have a bunch of nodes, but it does not make sense to put them in a cluster, so they are all individual
 
R3dM3at said:
I will admit i did not go through every page here yet, but if I need to re-install PDM, is there a file or somewhat to bring back all my nodes without having to re-add them 1x1? We have a bunch of nodes, but it does not make sense to put them in a cluster, so they are all individual
Click to expand...

Remote configuration is stored in /etc/proxmox-datacenter-manager/remtoes.cfg - if you transfer that one to a new PDM host, you should be good to go. Metric history is stored at /var/lib/proxmox-datacenter-manager/rrdb, in case you want to keep the statistics.
 
  • Like
Reactions: R3dM3at and Johannes S
Hi..

I installed the PDM.. but In the "allocation" section
It shows that I have:

CPU Cores assigned
197 running / 96 physical (6 total configured)

Memory assigned
272.865 GiB running / 439.907 GiB physical (12 GiB total configured)



Why does it only show 12GB and 6 configured CPU configured?
 

Attachments

  • runing.png
    runing.png
    16.1 KB · Views: 20
Hey friends!

This thing is amazing already; I'd like to throw my hat into the ring for a few feature requests, if I could...

1. The list of 'remotes' shows up in the drop-down on the left nav pane in chromium browsers (Opera in my case), but not Firefox. weird.

2. When a migration is underway, it'll show the task progress on the screen, but if you close it, I haven't found a way to get the status back (closest I've found is by looking at the bottom pane on the source node, but it doesn't seem to show anywhere in the datacenter UI if I close the overlay).

3. PBS integration - I know this is on the roadmap, but specifically, if I move a VM or container from one host to the other, the backup job fails because the object no longer exists on the host that's doing the backups. Is it possible to auto-update backup jobs so the backup jobs continue as they did before the move?

4. More of a PVE change, but I submit that it'd be nice to have a 'show fingerprint' button in the summary tab, the way PBS does, to streamline setups (I haven't had a problem once I knew to look at the cert, but still).

5. Would it be possible to add VMWare hosts to PDM for the purposes of migration? Obviously not a huge need, but it'd be a nice-to-have, since I've got a number of those on my docket =).

I think that's my wish list so far; thank you so much for making live migrations possible!!
 
The Search function at the top is case sensitive. Maybe intentional. Also returns the results of a node's storage (NFS/local/local-lvm) I don't see any of that information on the node in Remotes

1736450629164.png
 
Hello,
One suggestion, I have the ability to move between a standalone server and a cluster.
When moving from a standalone server to a cluster, there is no option on which server from the cluster to move the VM to.
Would, be nice to have such an option.

Best regards,
 
I think this product would serve really well as a container image, because other than some remote.cfg, there is not really much state to preserved. Any plans to support/release pdm as a container?
 
Last edited:
  • Like
Reactions: 4f1sh3r
While testing moving an LXC container from one system to another I noticed the following problems:
  1. In the logs the movement progress did not create a new line.
    1736844753306.png
  2. The VM ID is not the one set in the migration configuration window but remains the original one.
  3. Once I close the Task Viewer window I don't know where to find it in the pdm.
 
vaschthestampede said:
In the logs the movement progress did not create a new line.
Click to expand...
thanks for the report, i'll look into that

vaschthestampede said:
  1. The VM ID is not the one set in the migration configuration window but remains the original one.
Click to expand...
this is already fixed in git, not yet rolled out

vaschthestampede said:
  1. Once I close the Task Viewer window I don't know where to find it in the pdm.
Click to expand...
we're currently working on task tracking/caching for remotes so one can find them again ;)
 
  • Like
Reactions: vaschthestampede
I also encountered this other problem:
If you choose not to delete the VM, on the source node the VM can no longer be deleted (at least not easily but I haven't found how).

It was an LXC container and I used the command to unlock a VM. :eek:

Thanks for your work!
 
Last edited:
Hello,

Do you think it would be possible to use a proxy for internet output? The reason is that I have a dynamic IP, and on my Proxmox servers, I use a firewall to only allow access from one IP, which is a proxy IP.

Would it be possible to use a proxy for outgoing connections?

Thanks.
 
you could also create the token manually on the PVE side and set up the remote using it ;)
 
  • Like
Reactions: LeFred
kennysamuerto said:
Hello,

Do you think it would be possible to use a proxy for internet output? The reason is that I have a dynamic IP, and on my Proxmox servers, I use a firewall to only allow access from one IP, which is a proxy IP.

Would it be possible to use a proxy for outgoing connections?

Thanks.
Click to expand...

I have tried to configure the system-wide proxy on Debian, it works fine with the proxy's IP but the Proxmox filtered by IP does not work with this proxy IP.

Does anyone know if it's possible to force Proxmox Datacenter Manager to connect to the nodes using the proxy IP?
 
In some nodes, it allows adding them, and I can see real-time metrics on the VMs, but I do not get the usage history, neither of the node nor of the VMs. I find this error in the logs:

Jan 16 20:56:01 datacentermanager proxmox-datacenter-api[523]: failed to collect metrics for XXXXXX: api error (status = 501 Not Implemented): Method 'GET /cluster/metrics/export' not implemented
 
You must log in or register to reply here.
Top Bottom