The question is "FOR WHAT"? Just for running it?
I see many "node runners" but few have a good idea what really is for. When somebody say "I am going to run a node..." 1st thing they should ask themselves "What am I going to do with this node? For what I would run it?"
Please read more:
Listen to Darth.
Don't bother buying a bunch of dedicated hardware unless you've got a specific use case in mind that requires it.
Want to manage your cold storage funds? Running Core on a laptop is more than enough (you can keep the block data on an external drive, it doesn't need to stay on 24/7 -- once the initial block sync is done catching up on even a few weeks or months is quick).
Want to run a lightning node? Not worth it unless you run a business that accepts lightning payments or start an LSP which is a whole different level. For everyday lightning spending get a mobile node like Blixt, Zeus, Phoenix or even the latest Green wallet.
reply
there is a third reason to run a node - you want to work in the industry, build on lightning or study it. In that case go ahead.
while I agree with not running nodes for spending etc, and that most people will eventually lose money if they do, we really shouldn't be discouraging people who actually want to build proper setups instead of using completely unsuited hardware like raspis.
we as an industry need orders of magnitudes more experienced node runners and builders - we should be encouraging that while providing disclaimers of course.
what we should really be discouraging is people running nodes on shit hardware with implementations that are designed for people to lose funds
reply
I'm curious, what are implementations that are designed for people to lose funds?
I'm not a developer or builder, so my interest of running a node is purely personal and my intention is to add a few other services to the machine to make it worth running 24/7.
Right now I'm more leaning towards Ubuntu Server with a manual setup, simply because I like to have full control over what's going on and the available services to use.
reply
@M_affirmed the losing funds part was a bit cheeky but lnd architecture is not really built with robust deployments in mind as primary storage option for state is boltdb which you need to hack your way around to have any way of backing up sanely - hence all the raid/filesystem redundancy talk in node runner circles. but it has significantly more tooling out there than alternatives so ppl still go for it
security wise don't run other (at least publicly accessible) software on the same machine as your node, its calling for trouble. if you want to make better use of hardware throw some virtualization in the mix to isolate the services
reply
finally somebody is paying attention to the details...
reply