I’m setting up a new pioreactor with 64-bit version of the image to be able to run 64-bit apps on the leader. Is there a way I can take an existing leader and add it to the network of this new leader without having to reflash and lose the data/configs/plugins etc.?
- If you’re on the latest version, 25.9.x, there is a “Export” button on the Leader page to export the entire ~/.pioreactor/ folder. You can keep this as a backup. If not on the latest version, you can zip the .pioreactor folder, and then scp it off.
- On the new leader’s Inventory page, pretty sure if you just add the old leader via the “Add new Pioreactor” flow, it will just start acting like a worker. When you do this, the old config.ini is wiped, hence why I suggest exporting that data first. The old database, nor plugins, are not wiped.
Does that make sense?
Thanks! And does discovery happen over a tailscale network? Currently the old pioreactor isn’t being found
It should. If you’re having trouble, you can use the new leader’s CLI to be more precise with the address:
pio workers add --address <tailscale ip> -m <model-name> -v <model-version> <OLD_LEADER_NAME>
Great thanks! I’ll try that. Need to wait for the previous one to finish the current run in about an hour first 
Will report back
Seems to hang after running the command - guessing it shouldn’t take that long. It might be because the password isn’t the default one anymore - I log in with an ssh key. Is there any way to run it with debug or any thoughts about what the problem might be?
pio workers add --address 100.77.93.20 -m pioreactor_40ml -v 1.0 pioreactor01
pio logs:
2025-11-04T19:46:51+0000 [add_pioreactor] INFO Adding new pioreactor pioreactor01 to cluster.
Ok I think I fixed it by copying the ssh key from the new leader to the old leader. It’s added now