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.
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.
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.
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 I probably hadn’t so much as SSH’d into the node for 1-2 months, so clients were all out of date.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.
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.
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?