The mnemonic part is clear.
But how does one ensures the continuous operation of their LN node?
- How often or when are we supposed to backup the LN channel states? Or is it automatically backed up by Umbrel in some way?
- If the SSD were to break, replacing it and then re-downloading the entire blockchain again is likely going to take more than 24h and therefore is it a risk that someone may force close a channel on you? Does one needs to backup the SSD?
- Anything else?
It’s not recommended to try backup your channel states (channel.db). The issue with restoring from channel states is using an older state could cause funds in that channel to be forfeit and it’s difficult to guarantee you haven’t sent/received a transaction since the backup. It also can be a pretty big file for older/larger nodes.
Instead you should use static channel backup (channel.backup). You can download this from the lightning dashboard with the “…” next to “+ OPEN CHANNEL”. Instructions are in the FAQ on how to recover with SCB. Recovering with SCB requests the nodes you were connected to to close the channels and let you recover your funds.
For your questions:
- You should backup whenever a new channel opens and also before you update / do anything risky to your node.
- Force closing a channel typically gives a 2 week window (2016 blocks) for you come back and ensure they aren’t using a previous state. This should be enough time to get up and running again, but if not, you can also set up a watchtower to look for fraudulent closes while you are down.
- Future eltoo channels should allow for the correct channel state to be used if an old one is broadcasted, removing the need to forfeit all your funds.
Thank you so much!
So, just to be clear. An SCB (Static Channel Backup) won’t restore my channels, but it will help me recover funds by requesting channel closures. I will have to re-open the channels again, right?
No worries and, correct, you will be starting again.
So this is great to recover funds, but what are the options for ensuring redundancy of operations? ie I want to preserve the effort I put into building the reputation and channels, starting over is doable but not preferable.
Are there any recommended practices others have used for backing up the operation, both hardware and channel operations?
I think the best option is to have a RAID + USP setup, so hardware or power fail won’t affect your node. This is costly, but there is no way around it.
It would be cool to figure out a way to backup your channels files continously to the cloud, but even if we had this. You would still need new hardware ready to go from the backups if the original were to fail. So the first option is still much more preferrable.
What files are stored to the ssd and what files to the SD card? is it sufficient to just do permanent backups of ssd to get a FULL restore capability of the node including on and off chain transactions at the time the system crashes due to a hardware failure?
I learned the hard way not to rely on a single lightning node. I had a random hardware failure that resulted in having to spin up a new node, which meant having to close all my previous channels to recover my funds.
I’m currently exploring how to run a lightning node on my Rockpro64 cluster. Redundant everything.