16:53:33 s​yntheticbird:monero.social: Contrary to what has been said by telepathy from the moon, today's Cuprate meeting is **NOT** cancelled
16:53:44 s​yntheticbird:monero.social: every participants must be ready and in position to their battle stations
17:59:52 b​oog900:monero.social: !meeting
17:59:54 m​oo:monero.social: Starting meeting: https://github.com/monero-project/meta/issues/1312
17:59:58 b​oog900:monero.social: 1) greetings
18:05:34 s​yntheticbird:monero.social: Hi people
18:05:36 s​yntheticbird:monero.social: syn here
18:06:28 b​oog900:monero.social: 2) updates
18:06:30 s​yntheticbird:monero.social: there was a reference
18:09:23 b​oog900:monero.social: me: I did some work hardening RPC sync, making it more stable etc. Also last week I was reading up on some databases and it made me want to give fjall another go. I spent some time integrating it properly and it performs incredibly well.
18:15:59 h​into:monero.social: hello, sorry I'm late, update: still working on PoWER for ZMQ
18:16:26 b​oog900:monero.social: 3) Project: What is next for Cuprate?
18:19:16 b​oog900:monero.social: So the main thing I want to discuss today is how we feel about moving to fjall.
18:19:16 b​oog900:monero.social: The pros are obvious, it is much faster at syncing for every system. The more storage I/O constrained the better the improvement. My NVMe ssd goes from a little over 2hrs to sync to a little over 1hr. My HDD goes from days to weeks, to 1hr 20 mins.
18:21:34 h​into:monero.social: Are fjall DBs readable from multiple processes?
18:21:39 b​oog900:monero.social: The cons are that it is a less battle tested DB. And I don't think it supports multi-process access yet.
18:22:04 h​into:monero.social: How is the memory usage compared to LMDB?
18:22:11 b​oog900:monero.social: yeah I think it does if both are read only, but I am unsure about 1 writer 1 reader
18:23:51 b​oog900:monero.social: it is more, but not crazy amounts more. I don't have exact numbers. You can configure how much memory it uses.
18:24:48 b​oog900:monero.social: I am going to test a sync on a Pi 4 with a USB HDD soon, we'll see then if it can hold up.
18:26:18 b​oog900:monero.social: pining sgp_ as using fjall will break the block explorer magic made.
18:26:39 b​oog900:monero.social: pinging*
18:26:48 h​into:monero.social: Is there a public working branch? I can add test data as well
18:27:15 h​into:monero.social: e.g Pi 5 and M4 Mac mini
18:27:59 b​oog900:monero.social: there is a messy public branch for my testing, I wouldn't recommend you running it though as it is full of todos
18:28:51 b​oog900:monero.social: you _can_ run it but I literally built it out just to test sync
18:29:43 h​into:monero.social: If we do stick to a single DB I suggest removing the others such that we don't have to maintain them
18:29:55 h​into:monero.social: `cuprate-database` traits can then be more tailored to represent whichever DB is chosen
18:30:34 h​into:monero.social: (if we keep the traits around instead of directly using the DB)
18:30:47 b​oog900:monero.social: currently I am using fjall directly in cuprate-blockchain
18:30:52 s​gp_:monero.social: :'(
18:31:26 s​gp_:monero.social: Ideally I'd like to read it with a different program but I guess we could migrate to RPC or something if you're set on transitioning
18:33:44 b​oog900:monero.social: yeah it is annoying, I do think the performance gains are worth it though
18:36:17 b​oog900:monero.social: tbf you would still have access to the tapes database which store most of our data
18:36:33 b​oog900:monero.social: you would only need to fall back to RPC for some data, like hashes to indexes.
18:37:14 s​gp_:monero.social: don't skip those because of me then, so long as our thingy can be adapted
18:38:17 b​oog900:monero.social: hinto: do you have any objections at the moment ?
18:39:21 h​into:monero.social: No, but it should be tested on a few machines before final merge
18:41:11 b​oog900:monero.social: hinto do you have anything else you want to discuss?
18:44:36 s​gp_:monero.social: if one of you has some time to put together a basic gist of how we should use it going forward after the change, that would be great. Otherwise we'll figure it out
18:45:25 h​into:monero.social: We can end here
18:45:49 b​oog900:monero.social: I'll do that before we merge it to main
18:45:58 b​oog900:monero.social: thanks everyone!
18:46:03 b​oog900:monero.social: !meeting
18:46:06 m​oo:monero.social: - Logs: https://github.com/monero-project/meta/issues/1312#issuecomment-3687611153
18:46:06 m​oo:monero.social: - Next meeting: https://github.com/monero-project/meta/issues/1316
18:47:14 plowsof: 👍
18:47:24 b​oog900:monero.social: it turns out a 2GB Pi 4 can't build Cuprate :(
18:47:30 b​oog900:monero.social: it gets killed
18:50:24 s​yntheticbird:monero.social: build
18:50:26 s​yntheticbird:monero.social: you mean sync
18:50:35 b​oog900:monero.social: nah I mean build
18:50:46 s​yntheticbird:monero.social: no you don't
18:51:11 s​yntheticbird:monero.social: sorry gaslighting attempt
18:51:16 plowsof: increase swap and let er rip
18:51:35 s​yntheticbird:monero.social: this
20:54:17 b​oog900:monero.social: !cancel 1 Holidays
20:54:18 m​oo:monero.social: Command parse error: IncorrectParameter
20:54:32 b​oog900:monero.social: !cancel 1 "Holidays"
20:54:35 m​oo:monero.social: Canceled meeting(s):
20:54:36 m​oo:monero.social: - https://github.com/monero-project/meta/issues/1316
20:54:38 m​oo:monero.social: Next meeting: https://github.com/monero-project/meta/issues/1317
20:56:00 b​oog900:monero.social: hmm it created a meeting with the same date, I guess I was supposed to do that after the meeting?
21:44:28 h​into:monero.social: apparently I fixed that last year: https://github.com/Cuprate/moo/commit/024325d63b745d01c31b6ef902c75550311de3bf
21:44:47 h​into:monero.social: !status
21:44:48 m​oo:monero.social: moo v0.0.18, 3babb9799bee8440adec832446a1522ee6c0b78e (release), meeting: false, uptime: 4 months, 25 days, 8 hours, 4 minutes, 35 seconds
21:44:51 h​into:monero.social: !shutdown
21:44:52 m​oo:monero.social: Shutting down gracefully, uptime: 4 months, 25 days, 8 hours, 4 minutes, 40 seconds
21:51:16 h​into:monero.social: !status
21:51:18 m​oo:monero.social: moo v0.0.19, abea80235b3d9f86b5f425a85827c41e94b94793 (release), meeting: false, uptime: 28 seconds
21:51:45 h​into:monero.social: will the upgraded moo actually work this time? I guess we'll see in a year...