Cant add new node to cluster, error400

[Mark van de Louw]

Hello all,

At this moment, I am trying to add a new node (fresh install) to our excisting cluster.

The node itself is on PVE6.0-4 and the other nodes are still on 5.4-13. i Need to add this node to miigrate the VM's towards, so I can update the other hosts.

I get the following error when trying to add the node:
Please enter superuser (root) password for 'HIDDEN':
Password for root@HIDDEN: ********************
Establishing API connection with host 'HIDDEN'
The authenticity of host 'HIDDEN' can't be established.
X509 SHA256 key fingerprint is HIDDEN.
Are you sure you want to continue connecting (yes/no)? Login succeeded.
Request addition of this node

TASK ERROR: 400 Parameter verification failed. link0: property is not defined in schema and the schema does not allow additional properties



Does anyone have a idea what can cause this issue?
Within the gui, i pasted the cluster information and selected the correct NIC/IP in the Link 0 dropdown.


Kind regards

 

[tom]

The node itself is on PVE6.0-4 and the other nodes are still on 5.4-13.

this is not possible.

see also:

https://pve.proxmox.com/wiki/Upgrade_from_5.x_to_6.0

 

[janos]

this is not possible.

see also:

https://pve.proxmox.com/wiki/Upgrade_from_5.x_to_6.0

Even when pve5 nodes upgraded to corosync 3?

 

[oguz]

Even when pve5 nodes upgraded to corosync 3?

no. you need to have all nodes on the same PVE version.

 

[harumscarum70]

This..... Saddens me.

Could I qm export the VM's from my 5.3 nodes... and qm import them into a fresh new 6.1 install? Hmmmm... I guess the challenge... is the shared storage... because I can't share storage between 2 disparate clusters... correct? So if I don't have sufficient space to run the server local... I could qm export the image somewhere.. like NFS I suppose... and then qm import back onto my shared iscsi array? Sorry... lots in here to dissect perhaps.

 

[oguz]

Could I qm export the VM's from my 5.3 nodes... and qm import them into a fresh new 6.1 install?

yes you can

I could qm export the image somewhere.. like NFS I suppose... and then qm import back onto my shared iscsi array?

yes

 

[harumscarum70]

thanks... actually did it with vzdump... any thoughts on which might be quicker? more reliable?

 

[oguz]

any thoughts on which might be quicker?

you can play around with the compression options while backing up to see what's faster

more reliable?

what is unreliable with vzdump?

 

[harumscarum70]

"what is unreliable with vzdump?"

Haha... I don't actually know... my first experience using it and seems ok so far. My network is only 1 gig here though so this 1.5 TB server has been vzdumping for 7 hours now.....

Another quick question... thank you so much by the way...

If I literally just shut down VM's and attach my new proxmox 6.1 host to the iscsi array... would I be able to simply find the disks? Instead of export import? Or is that simply not possible since it will be a new server outside of the proxmox 5 cluster... and therfore it will not understand the filesystem and see the disk???

Meaning literally... my only real options if i want exact replicas of my servers moved to new 6.1 is to export (vzdump likely) and import (qm restore) to new.... seems logical... since it will bring the conf with the restore.

Thanks again... so much... very helpful

 

[oguz]

you could use the entries from your old cluster's /etc/pve/storage.cfg to use the iscsi on the new cluster

 

[oguz]

you can check for example here[0]

that page is for PVE 3.0, but the process is the same

[0]: https://pve.proxmox.com/wiki/Upgrade_from_3.x_to_4.0#Bypassing_Backup_and_Restore

 

[harumscarum70]

oh.... dude... if this works.... you are awesome. I can totally just save the important files... the vmid conf files... network interfaces... storage... reinstall my 2 remaining servers with 6.1.... slap the config files on them... and bang... good to go?

So... TECHNICALLY... what I'm hearing you say is even though it is shared storage... I could right now... point my 6.1 server to a currently hosted 5.4 VM. By shutting the vm down on 5.4.... Take the vmid.conf file off the 5.4... and restore on the 6.1. To avoid the entire vzdump restore process?????????

So it's actually ok for me to use shared storage between clusters in that use case at least... short term...

I could move VM's a few at a time to the 6.1 host... upgrade the others... then migrate them where I want them.