austonst avatar

austonst

@austonst@kbin.social
austonst,
austonst avatar

Anyone have experience with Lodestar? Running a relay (Aestus) requires CL clients with a little bit of different code to implement relay-specific behavior. Currently only Lighthouse and Prysm have the appropriate changes, and Prysm has been giving us some weird issues with very high CPU and RAM usage, so we're a little Lighthouse-heavy. I love Lighthouse, but we'd like to diversify a bit. Lodestar has an open issue to implement the necessary changes, which is promising.

But Lodestar is probably the only CL client that I've never worked with. I know it's probably easy enough to figure out how to build a docker image for it, and look up the usual set of command line options, but I'm trying to find whatever excuses I can to post to kbin. So, any thoughts on Lodestar?

austonst,
austonst avatar

Continuing yesterday's node maintenance saga. In the end it looks like I didn't even hit 48 hours of downtime; Besu can sync from scratch pretty quickly and then everything went right back to normal. I've had much longer outages than this over the years, definitely learned to roll with the punches a little.

I did miss a block proposal, which stings a little bit, but I have enough validators that it's not like I missed my only chance this month.

austonst,
austonst avatar

I can't let myself play OpenTTD anymore. It runs too big a risk of annihilating my sleep schedule. There's always some city that could really use some bus route optimization.

austonst,
austonst avatar

My home staking Besu decided to die on me, got its database corrupted, and is now costing me $$$ while it resyncs from scratch. It's almost completely my fault, so I'm not complaining. I probably hadn't so much as SSH'd into the node for 1-2 months, so clients were all out of date. On top of that, my phone started giving me notifications that Besu was having some periodic downtime; I think if I don't restart the node for a long time I get accumulated memory leaks of some sort that start giving Besu OOM errors. And while it does auto-restart, at some point those crashes are going to corrupt my DB. But I took a few days to get around to it, and by then it was too late.

Don't be lazy, maintain your nodes.

austonst,
austonst avatar

Eh, probably just need to be better about staying on top of client updates. If every time there's a Besu or Lighthouse release, I check in, do a full system update, and restart clients as needed, it'll probably be frequent enough to keep everything running smoothly.

austonst,
austonst avatar

MEV-Boost Community Call #4 was yesterday, nothing particularly spicy but a lot of good updates on relay development.

You can find the agenda here and the recording here.

austonst,
austonst avatar

I briefly checked out kbin a few days ago, but took a deeper dive today... and I'm kind of on board now. Still don't quite get the relationships between magazines, threads, and microblogs, and unsure how federated content from platforms with a different structure and UX (say, Mastodon) will map onto kbin. But my big realization today is that I don't have to understand all that to get started!

If you forget about federation and microblogs for a bit, kbin.social seems like a basic but serviceable Reddit clone. You can find various nascent communities popping up that seem like they function similarly enough as they do on Reddit. Daily threads on @ethfinance, episode discussion threads on @anime, strong political biases already taking root in @news, etc. I can understand all that without caring what goes on behind the scenes. And I can explore the larger fediverse at my own pace.

The big challenge is just that network effects are too stupidly strong. Makes it way too hard to unseat incumbents. If I ever picked up a new hobby or video game, or wanted to see what the experts on home ketchup production thought about how different brands of sugar affect the mouthfeel of the final product, I could be sure to find a thriving community on Reddit discussing just that. It's tough to give that up. But right now the promises of federation and the hopes of creating an ecosystem more resilient to large-scale corporate BS are sounding pretty good to me.

But to keep things on topic, I wish I could sign in with Ethereum and have my profile and data controlled by my Ethereum address. Make identity intrinsically transferable across instances/servers, have native support for POAPs/NFTs, and possibly build a lot of identity tools around ZK attestations of some sort. I wonder if ActivityPub is flexible enough that a web3-oriented frontend could be developed to provide that kind of UX while largely relying on other instances for communities and content.

  • All
  • Subscribed
  • Moderated
  • Favorites
  • anitta
  • mdbf
  • magazineikmin
  • InstantRegret
  • hgfsjryuu7
  • Durango
  • Youngstown
  • slotface
  • everett
  • thenastyranch
  • rosin
  • kavyap
  • khanakhh
  • PowerRangers
  • Leos
  • DreamBathrooms
  • vwfavf
  • ethstaker
  • tacticalgear
  • cubers
  • ngwrru68w68
  • modclub
  • cisconetworking
  • osvaldo12
  • GTA5RPClips
  • normalnudes
  • tester
  • provamag3
  • All magazines