Why I started and why I stopped actively running my own Bitcoin Lightning node.
I run 3 nodes:
An upgraded Umbrel Home with the Bitcoin Core and Lightning Node, this is a full node and has several channels and if I keep it balanced, which I don’t bother to do now, could earn routing fees. I also run a full Liquid blockchain for any L-BTC or L-USDT that I have.
An upgraded RPi 5 8GB RAM with 2TB SSD, this was a full node, but Lightning crashed several times and I lost 0.2 BTC in channels due to not backing up sufficiently. I don’t trust this to hold liquidity, but I do run it for other tasks, such as running the Monero blockchain and experimenting with password managers, BitTorrent clients, server monitoring and password managers.
Start9 node running on a Lenovo NUC with 32GB RAM & a 2TB SSD. This is currently powered down.
When I first setup my 2 Umbrel nodes, originally RPIs with 4GB + 2 TB SSD & 8GB + 1TB SSD respectively, I used the Lightning node itself to open channels using https://amboss.space/ and https://lightningnetwork.plus/ to find suitable nodes.
This was easy and cheap at the time. I remember setting up several triangle nodes to small capacity nodes on my first Umbrel and several large capacity channels to large public nodes on my second Umbrel. I also setup large channels between my own two nodes so I could route payments across my own network. I was experimenting heavily.
To rebalance, I installed RTL, Ride the Lightning and Thunderhub, but quickly realised that RTL was the tool for me. I also installed an auto rebalancing tool called LNDg, which I spent many months watching YouTube videos and reading articles in a failed attempt to master.
During this period I was spending at least 8 hours a day in channel management and learning, I had around 0.5 BTC in open channels on each node.
At the peak, I was maybe earning a few thousand Sats a day for my efforts.
It quickly became very apparent that this wasn’t sustainable and I could never earn enough to justify the time spent in manual management and so I started looking for automation tools that could free up my time.
There were some tools emerging, but they required manual configuration, not walled Umbrel nodes to work on, and they required coding skills to configure, which I lacked.
I toyed with the idea of resurrecting my coding skills, but as I hadn’t coded since the 3GL days of the late 1980’s to early 1990’s I realised that my heart wasn’t in it and I couldn’t find enough enthusiasm to do this, so I let my nodes languish.
At this point, I discovered Start9, which appeared promising at first, but when I actually installed it on a dedicated NUC I had purchased on eBay, I realised it was even more complex and was even more of a walled garden, so I never bothered to load any liquidity on this and ran it for a year or so before switching it off a few months ago.
I am waiting for the full release of Start9 which promises a much better UI and support for VPNs, which I require.
In the meantime, I still run one full node and have significant liquidity on it, but I mostly use it for my own payments. I do sometimes rebalance the channels and watch the fees roll in for a few hours until the channels are unbalanced again.
I would start this properly if I thought it was financially viable, but even Alex Bosworth, who runs one of the largest nodes on the network admits it doesn’t pay enough to make it worth the time. He is on X and you can follow him for much more detailed advice here: