One backup job to multiple nodes...

G

Gilberto Ferreira

Renowned Member
Aug 21, 2014
444
53
93
Joinville, Santa Catarina, Brazil, Brazil
Hi

I have a cluster with 5 nodes: pve01-dc3, pve02-dc1,pve01-dc1,pve03-dc2,pve04-dc2
, but I am using only 4 nodes.
So I created a job via web gui, but I noticed that I can't choose more than one node, meaning I have to create a separate jobs for each node.
Is that right?
I tried to do this:
vzdump: backup-1fd712c4-51a3
schedule 00:30
all 1
enabled 1
fleecing 0
mode snapshot
notes-template {{guestname}}
notification-mode notification-system
repeat-missed 0
storage bkp
nodes pve02-dc1,pve01-dc1,pve03-dc2,pve04-dc2

But I still have this error message:

TASK ERROR: could not activate storage 'bkp': storage 'bkp' is not available on node 'pve01-dc3'

I'll appreciated for any advice.

Thanks.
 
Last edited:
You can do that with the backup. PVE is telling you something is wrong when accessing bkp on node pve01-dc3.

Please post the output from this command:

cat /etc/pve/storage.cfg | awk 'BEGIN{RS=""} /^bkp:/'

It will output your cluster storage configuration and filter for the bkp storage configuration.
 
weehooey-bh said:
You can do that with the backup. PVE is telling you something is wrong when accessing bkp on node pve01-dc3.

Please post the output from this command:

cat /etc/pve/storage.cfg | awk 'BEGIN{RS=""} /^bkp:/'

It will output your cluster storage configuration and filter for the bkp storage configuration.
Click to expand...
Of course is telling me that something is wrong, because I do not have any VM running in pve01-dc3, hence I do not need backup any vm from that node.
So, I need something to exclude that node or via WEB UI allows me to choose onlyu the nodes I want job backup. But I can't do that.
Simple like that.

Anyway, thanks for you considerations.
 
Okay. Then provide the full output of cat /etc/pve/storage.cfg

If you cannot select the nodes in the Create: Backup Job dialog box, then it is likely you have an issue with your storage configuration.
 
weehooey-bh said:
Okay. Then provide the full output of cat /etc/pve/storage.cfg
Click to expand...
Here we go:
dir: local
path /var/lib/vz
content import,iso,images
preallocation off
prune-backups keep-all=1
shared 1

rbd: VMS
content images,rootdir
krbd 0
nodes pve01-jdm,pve04-sta,pve03-sta,pve02-jdm
pool VMS

pbs: bkp
datastore backup
server SERVER_IP
content backup
fingerprint bb:4d:79:e3:ac:75:5d:2c:04:f2:e4:3f:a1:22:eb:8c:6f:41:e0:b8:a1:44:f6:40:d2:0e:05:1a:2a:59:f8:89
nodes pve02-jdm,pve01-jdm,pve03-sta,pve04-sta
prune-backups keep-all=1
username admin@pbs


As I said, I had choosed to have pbs: bkp only on the nodes I have running VMS.
 
Based on your config "bkp" is a PBS server, which means you more than likely added it via the Datacenter > Storage setting - That should have replicated to all the nodes in the cluster. Even if you don't run VMs on that node, you'd still want that storage attached it.
 
  • Looking at your storage.cfg, it appears -dc1 and -dc2 in the PVE hostname corresponds to -jdm and -sta
  • Node pve01-dc3 is not included in the storage definition for bkp: nodes pve02-jdm,pve01-jdm,pve03-sta,pve04-sta
  • Your backup job vzdump: backup-1fd712c4-51a3 is set to run on all your nodes: all 1
So you need to either:
  • Allow bkp to be available on all nodes in the storage definition, or,
  • Set the backup job only to include the target nodes
To leave bkp as is (which is what I believe you want) and to select the target nodes in the web GUI:
  1. Click the "Node" drop-down box
  2. Click on each of the nodes you want to include (they will turn blue when selected)

backup_dialog.png
 
I was about to add what BH said. You probably need to limit the backup job to the nodes that the storage is attached too. Otherwise it will have an issue with the backup process.
 
  • Like
Reactions: weehooey-bh
weehooey-bh said:
  • Looking at your storage.cfg, it appears -dc1 and -dc2 in the PVE hostname corresponds to -jdm and -sta
  • Node pve01-dc3 is not included in the storage definition for bkp: nodes pve02-jdm,pve01-jdm,pve03-sta,pve04-sta
  • Your backup job vzdump: backup-1fd712c4-51a3 is set to run on all your nodes: all 1
So you need to either:
  • Allow bkp to be available on all nodes in the storage definition, or,
  • Set the backup job only to include the target nodes
To leave bkp as is (which is what I believe you want) and to select the target nodes in the web GUI:
  1. Click the "Node" drop-down box
  2. Click on each of the nodes you want to include (they will turn blue when selected)

View attachment 88726
Click to expand...
My God man... I told you guys right there on the top, I can not select multiple nodes on the backup task!

Read it again: https://forum.proxmox.com/threads/one-backup-job-to-multiple-nodes.169126/#post-787764

Here is a little video showing this:

And yes I tried to do manualy like this:
Code: 
nodes pve02-dc1,pve01-dc1,pve03-dc2,pve04-dc2
In side the jobs.cfg and there is no effect.
 

Attachments

  • vokoscreenNG-2025-08-01_12-58-34.gif
    vokoscreenNG-2025-08-01_12-58-34.gif
    741.8 KB · Views: 6
Last edited:
Gilberto Ferreira said:
My God man... I told you guys right there on the top, I can not select multiple nodes on the backup task!

Read it again: https://forum.proxmox.com/threads/one-backup-job-to-multiple-nodes.169126/#post-787764

Here is a little video showing this:

And yes I tried to do manualy like this:

In side the jobs.cfg and there is no effect.
Click to expand...

Remove the 'nodes pve02-jdm,pve01-jdm,pve03-sta,pve04-sta' entry from your storage.cfg, you're limiting a cluster wide storage, so you can no longer select cluster wide nodes.

See if you then select which nodes the backup job can run from.

And not to get entirely off the subject, if that node is just used for CEPH, why not just install CEPH on a Debian install and add it to the CEPH storage cluster, but not the Proxmox cluster? You can have the two separate.
 
PwrBank said:
Remove the 'nodes pve02-jdm,pve01-jdm,pve03-sta,pve04-sta' entry from your storage.cfg, you're limiting a cluster wide storage, so you can no longer select cluster wide nodes.
Click to expand...
And I do that, with VMS RBD storage, in order to prevent one to try to migrate VM to that external node, because, as I said before, it's there just to ceph witness.
 
Last edited:
PwrBank said:
Remove the 'nodes pve02-jdm,pve01-jdm,pve03-sta,pve04-sta' entry from your storage.cfg, you're limiting a cluster wide storage, so you can no longer select cluster wide nodes.

See if you then select which nodes the backup job can run from.

And not to get entirely off the subject, if that node is just used for CEPH, why not just install CEPH on a Debian install and add it to the CEPH storage cluster, but not the Proxmox cluster? You can have the two separate.
Click to expand...
Ok. Now the PBS storage appears in all nodes, like so:
Code: 
pbs: bkp
        datastore backup
        server 172.16.250.206
        content backup
        fingerprint bb:4d:79:e3:ac:75:5d:2c:04:f2:e4:3f:a1:22:eb:8c:6f:41:e0:b8:a1:44:f6:40:d2:0e:05:1a:2a:59:f8:89
        prune-backups keep-all=1
        username admin@pbs

But even so, I can not select the nodes I want... As you can see in the video.
 
Gilberto Ferreira said:
And I do that, in order to prevent one to try to migrate VM to that external node, because, as I said before, it's there just to ceph witness.
Click to expand...
...But it's a backup storage, not a normal storage. VMs will migrate to it unless you create a group for the compute nodes.


If you're trying to prevent VMs from moving to this node you need to create a group in the datacenter > HA section
1754068948441.png
Then VMs will be limited to what they can and cannot be running on.
 

Attachments

  • 1754068889194.png
    1754068889194.png
    16 KB · Views: 0
Last edited:
PwrBank said:
...But it's a backup storage, not a normal storage. VMs will migrate to it unless you create a pool for the compute nodes.
Click to expand...
yeah yeah... I had edited my comments, like that:
And I do that, with VMS RBD storage, in order to prevent one to try to migrate VM to that external node, because, as I said before, it's there just to ceph witness.

I was think you were talk about RBD storage.
 
PwrBank said:
...But it's a backup storage, not a normal storage. VMs will migrate to it unless you create a group for the compute nodes.


If you're trying to prevent VMs from moving to this node you need to create a group in the datacenter > HA section
View attachment 88737
Then VMs will be limited to what they can and cannot be running on.
Click to expand...
yes... I alread had set this HA group... This is not my complain...
 
Gilberto Ferreira said:
yes... I alread had set this HA group... This is not my complain...
Click to expand...
You're saying you do not want VMs to run on that node, correct? If you already have that set in the HA settings then that is what will happen.

Now, I'm not sure why you're saying you don't want the backup storage to be attached to that node, because, again, nothing will run on it anyway, so it does not matter if it's attached or not. It will not try to backup anything to the storage, because there's nothing back up to it. There is no harm in having this backup listed attached to that node, unless for some reason you're worried that someone will restore the VM to that node.
 
You must log in or register to reply here.
Top Bottom