Those who know C++, how does bitcoind (Bitcoin Core) read/write data from/to disk? I'd love to tweak certain settings for my public node. Is bitcoind using specific page sizes? Please answer here, if possible:
New hardware for the nostr relay (among other things) ftp.halifax.rwth-aachen.de has been shipped! I'm excited, playing around with expensive hardware is a lot of fun :)
256 GByte RAM, 19x 4 TByte HDDs, 3x 800 GByte SSDs, and at least 2x 10 GBit/sec of redundant Internet connectivity
I'm pretty confident ftp.halifax.rwth-aachen.de will receive a hardware upgrade soon, thanks to a sponsor I'll name once it is official. The nostr relay (wss://ftp.halifax.rwth-aachen.de/nostr) will also benefit from this!
~75 TByte HDD disk space
several TByte SSD disk space (for caching)
192 GByte RAM (maybe 256 GByte, we'll see!)
at least 20 GBit/sec internet connectivity
7 years of warranty with on-site support
As before, the server will be hosted at RWTH Aachen University (sponsoring power, cooling, network).
This happened again, twice. I requested 2 sat/vByte, next block is at around 3-4 sat/vByte. Both peers wanted to pay 30 sat/vByte instead. Their loss...
View quoted note →
I'm lucky to have several large (~1 BTC and more) channels and enough inbound liquidity on my #LightningNetwork node. Because of this, I also see large forward requests. Some succeed, some fail.
However, it's always saddening when my node has to reject a 0.9 BTC forward request with lots of sats attached as fees I could claim. I simply don't have that much liquidity.
I wonder how this wrks out for other node operators - Alex Bosworth with his yalls nodes as an example.
It'd be nice for LSPs/wallets to split larger payments (MPP), possibly with #PickhardtPayments. That way, my node could at least serve a (large?) fraction of the total amount. But, who knows, maybe those 0.9 BTC in my example above are just 1% of the total amount :)
