I thought I had read that PDM would facilitate node to node transfers but I installed and kicked the tires and I see a migrate option but apparently that only works in cluster environments. When I select the migrate no other nodes come up so there is no way to complete. Is it possible and I am missing the option and if not is it going to be available in the future, thanks.
Node to node migration without a cluster
- Thread starter bishoptf
- Start date
hi -> no, you don't need a cluster - you can migrate from source "remote" to target "remote" - where "remote" is a single PVE host
https://pdm.proxmox.com/docs/remotes.html#
https://pdm.proxmox.com/docs/remotes.html#
Is it possible you are missing the step that you need to click the "Target Remote" pull-down in the Migrate menu and change to the other remote before selecting the target node?
Yeah Ive tried that, one node show up in the top but in the bottom pull down no nodes are available for selection. I have 2 nodes that I have added and they show up fine, both running 9.x version and the latest PDM version. Why I asked since no nodes show up in the bottom pull down...
Will take some screenshots when I am able this afternoon...
Last edited:
You can also use the pct remote-migrate and qm remote-migrate command line tools:
https://www.thomas-krenn.com/en/wiki/Proxmox_Remote_Migration
https://pve.proxmox.com/pve-docs/pct.1.html
https://pve.proxmox.com/pve-docs/qm.1.html
You could try if this works (you might want to create small test vms or lxcs so nothing important is lost if something goes wrong) because then it should work in the PDM too.
https://www.thomas-krenn.com/en/wiki/Proxmox_Remote_Migration
https://pve.proxmox.com/pve-docs/pct.1.html
https://pve.proxmox.com/pve-docs/qm.1.html
You could try if this works (you might want to create small test vms or lxcs so nothing important is lost if something goes wrong) because then it should work in the PDM too.
Like I thought I stated previously its not a cluster, just 2 individual nodes....
Can you post a screenshot of your migration screen and your configuration?
My migration screen for some vm looks like this, you can see that I would be able to migrate this VM from the source "pvenas" to the remote "pve-cluster"
I have two remotes, one being a cluster with two nodes (pve-cluster) and the other one a single-node install (pvenas):
My migration screen for some vm looks like this, you can see that I would be able to migrate this VM from the source "pvenas" to the remote "pve-cluster"
I have two remotes, one being a cluster with two nodes (pve-cluster) and the other one a single-node install (pvenas):
Looks like I wasn't seeing the forrest for the trees, but the lightbulb went off, I just wasn't reading the screen properly.....your screenshots helped me finally get it. I have been backing them up and then restoring them and its been working well but trying to save a step and speed things up on my migration. I will try one out and see how it works compared to the backup, I assume its just doing a zfs sync or something like that, I am not doing live migrations, power is off and just trying to move them around for the VMware migration.
Thanks again, sorry for being so daft, lol.
You are welcome 
To be honest I'm not sure how the migration actually work but I'm quite sure it's NOT doing zfs sync since it also works with other storage backends (like Ceph, NFS, LVM etc), so they must use different storage-agnostic mechanisms. This is a kind of double-edged sword though: On one hand it doesn't need zfs but is storage agnostic. This is quite important for large-scale environments who use Ceph in a hyper-converged setup, a NetApp Filer attached with NFS or a classical storage-area-network (for these you usually need LVM) or want to migrate from one type to another (e.G. from a Cluster with Ceph storage to a single-node with ZFS). On the other hand this means that it will propably be slower than an incremental zfs replication.
It would be a nice feature (no idea whether this is already planned or not) to use storage-native sync mechanisms like zfs send/receive if remote and source have the same storage backend type.
To be honest I'm not sure how the migration actually work but I'm quite sure it's NOT doing zfs sync since it also works with other storage backends (like Ceph, NFS, LVM etc), so they must use different storage-agnostic mechanisms. This is a kind of double-edged sword though: On one hand it doesn't need zfs but is storage agnostic. This is quite important for large-scale environments who use Ceph in a hyper-converged setup, a NetApp Filer attached with NFS or a classical storage-area-network (for these you usually need LVM) or want to migrate from one type to another (e.G. from a Cluster with Ceph storage to a single-node with ZFS). On the other hand this means that it will propably be slower than an incremental zfs replication. It would be a nice feature (no idea whether this is already planned or not) to use storage-native sync mechanisms like zfs send/receive if remote and source have the same storage backend type.
Last edited: