10:01:57 parazyd: Hi, was sent here from #monero-dev. Wondering if there is a practical usage example of the Tapes db somewhere?
10:02:00 parazyd: I'd like to try it out
10:16:29 plowsof: boog900
10:21:21 b​oog900:monero.social: Here is my WIP branch here for cuprate: https://github.com/Cuprate/cuprate/tree/fjall-wip/storage/blockchain/src
10:21:22 b​oog900:monero.social: Opening the tapes is done in `free.rs`.
10:24:23 b​oog900:monero.social: The tapes API should be pretty simple, we do need some docs though.
10:24:54 parazyd: ah cool. So is this like a backend for fjall db?
10:26:17 b​oog900:monero.social: No, we use both fjall and the tapes as separate databases
10:27:10 parazyd: oh ok, gotcha
10:27:17 parazyd: Thanks, I'll read it :)
10:27:17 b​oog900:monero.social: Its just we can't store some data in tapes
10:27:25 parazyd: Such as?
10:28:49 b​oog900:monero.social: On example is key images as they need to be sorted.
10:30:13 b​oog900:monero.social: Any data really that needs to be sorted by something other than the order they are put in the chain.
10:30:52 parazyd: I see, but Tapes could handle pure blocks and txs, right?
10:31:10 parazyd: Assuming there is an order to the txs in the block, but it's serialized anyway
10:34:44 b​oog900:monero.social: Yes but for blob tapes you need to store the index into the tape for your data elsewhere.
10:35:41 b​oog900:monero.social: For our block/tx blob tape we store the index in a fixed sized block/tx info tape
10:36:17 parazyd: Yes that makes sense, thanks
17:06:00 b​oog900:monero.social: Meeting here in under an hour
17:11:44 m​axlpm:tchncs.de: Hey everyone, I'm having trouble with syncing the Monero blockchain to my HDD because it's been a week and I'm only at 29 GB out of 100 GB. I've heard that cuprate is much faster, can I sync using cuprate and then use the same blockchain file with monerod?
17:18:16 m​axlpm:tchncs.de: Ah, in the docs it says that the database is not compatible with monerod. Sorry, I'll RTFM next time (probably)
17:24:40 b​oog900:monero.social: It'll only get slower, I wouldn't run monerod on a HDD
17:27:38 b​oog900:monero.social: Cuprate will be a lot better on HDDs, and hopefully we should have wallet support with our RPC soon
17:59:45 b​oog900:monero.social: !meeting
17:59:47 m​oo:monero.social: Starting meeting: https://github.com/monero-project/meta/issues/1327
18:00:10 b​oog900:monero.social: 1) greetings
18:00:14 Cindy: hi
18:02:13 h​into:monero.social: ehllo
18:02:42 b​oog900:monero.social: 2) updates
18:04:56 b​oog900:monero.social: Me: I added some proptest's state machine tests to the tapes database, which caught some bugs. I also began the process of removing our database abstraction and moving the txpool over to fjall
18:05:24 h​into:monero.social: me: mostly finished PoWER, I think I'll leave some notes, open for review this week and fix p2p merge conflicts later; maybe the Cuprate version could be opened now since it is just the main crate
18:06:44 b​oog900:monero.social: > Project: What is next for Cuprate?
18:06:49 b​oog900:monero.social: 3. Project: What is next for Cuprate?
18:07:23 b​oog900:monero.social: hinto: would you be against a 4 GB minimum RAM requirement for Cuprate?
18:08:14 b​oog900:monero.social: I want to add some dynamic cache settings based on the amount of RAM we are told the system has
18:08:48 h​into:monero.social: I might be remembering wrong but won't it be closer to that for post-FCMP monerod?
18:09:49 b​oog900:monero.social: probably, someone did sync after a few issues with 2 GB on stressnet though IIRC
18:09:54 Cindy: i have a question, does the memory requirement include memory overhead from the OS?
18:12:10 h​into:monero.social: IIRC the most purchased pi 5 models were 8gb, if we are using that as a benchmark
18:12:53 h​into:monero.social: although it does mean the low-cost VPS plans are no longer viable, those usually start at 1/2GB
18:13:43 b​oog900:monero.social: yeah, a 4GB system is the minimum not 4GB allocated to Cuprate, I plan to make the default setting to use 50% of system RAM upto a certain amount for each sub-system.
18:14:18 Cindy: i see
18:14:28 b​oog900:monero.social: it wouldn't be a hard minimum but it gives us a lower bound on maximum RAM usage + it sets expectations for people who try lower than 4 GB
18:14:50 b​oog900:monero.social: as I think even now for Cuprate and monerod that is probably too low
18:16:05 b​oog900:monero.social: the hdd ones are now available though :D
18:16:10 h​into:monero.social: I've had `cuprated` running well up until now with a 1GB cap btw
18:16:24 b​oog900:monero.social: wow really 👀
18:16:42 b​oog900:monero.social: exceeding what I thought it could do lol
18:16:56 b​oog900:monero.social: did you sync with 1 GB?
18:17:18 h​into:monero.social: although it eats ram if you give it more, seems to peak at 3.5gb, I have it capped at 2gb currently
18:17:34 b​oog900:monero.social: that would be LMDB :)
18:17:36 h​into:monero.social: yes
18:19:37 h​into:monero.social: 1.3gb resident, 0.7 shared, so possibly other things too
18:21:36 s​yntheticbird:monero.social: Hi
18:23:04 b​oog900:monero.social: Ok that is interesting, I guess no need to agree to 4 GB today we can see how the new database performs on everyones systems first. One of the "problems" with fjall and the tapes is you need to set the cache sizes yourself
18:23:18 b​oog900:monero.social: you can't rely on the system to eat all the RAM you give it
18:23:37 b​oog900:monero.social: blessing and a curse as we do get more control over what is cached.
18:26:18 b​oog900:monero.social: anything anyone else wants to discuss today?
18:26:50 h​into:monero.social: Should I open https://github.com/Cuprate/cuprate/pull/568 for review? or wait until upstream is merged?
18:27:03 g​ingeropolous:monero.social: are peeps mining on it yet?
18:27:41 b​oog900:monero.social: I would wait
18:29:25 b​oog900:monero.social: no and out of every activity you can do on Cuprate that is probably the least recommended. Mining on Cuprate could make chain splits worse if they were to happen.
18:29:40 Cindy: i appreciate the modularity of cuprate
18:29:54 Cindy: i use some parts of it in my own program
18:30:12 Cindy: i know i said that before :P
18:30:44 b​oog900:monero.social: so this should probably wait for 1.0, at least for any big mining operations
18:31:21 b​oog900:monero.social: Make sure to let us know when you make it public :)
18:31:41 Cindy: i have
18:31:53 b​oog900:monero.social: link?
18:32:10 Cindy: https://codeberg.org/techmetx11/xmrgift
18:33:58 b​oog900:monero.social: I can't see Cuprate being used :(
18:33:58 Cindy: i've only tested it on stagenet
18:34:02 b​oog900:monero.social: only monero-oxide
18:34:12 Cindy: i remember cuprate using monero-oxide though :o
18:34:19 Cindy: maybe i misremembered
18:34:37 s​yntheticbird:monero.social: because vikings drink water and I drink water doesn't mean i'm a viking
18:34:48 Cindy: yeah true
18:36:04 b​oog900:monero.social: FWIW here is cuprate: https://github.com/Cuprate/cuprate
18:36:05 b​oog900:monero.social: It's basically node parts, rather than monero-oxdie which is wallet/underlying types parts
18:36:25 b​oog900:monero.social: I think we can end here, thanks everyone!
18:36:34 h​into:monero.social: thanks
18:36:40 b​oog900:monero.social: !meeting
18:36:41 Cindy: well looks like i got confused then!
18:36:45 m​oo:monero.social: - Logs: https://github.com/monero-project/meta/issues/1327#issuecomment-3774380023
18:36:46 m​oo:monero.social: - Next meeting: https://github.com/monero-project/meta/issues/1329
18:37:14 b​oog900:monero.social: lol its fine, hopefully you can find something to use the _real_ Cuprate in :)
18:59:42 Cindy: can cuprate be used in a no-std environment?
18:59:54 b​oog900:monero.social: some of it can
19:01:47 Cindy: it's not like it's feasible to run cuprate in an embedded environment
19:01:51 Cindy: due to the massive RAM usage
19:02:06 Cindy: maybe just monero-oxide
19:13:45 b​oog900:monero.social: You could use just the p2p libs for something, but yeah running full cuprate is probably not a good idea.