00:00:13 selsta, are you sure? 00:00:14 My thinking is that even if 100% of the pre-flood tx volume (~100kB per block) is paying higher fees, that volume is not enough to consistantly fill 300+ kB blocks. By the time miners build a 300 kB block, there would rarely be any high fee txs left to push against the penalty median. 00:00:59 I think only the flood txs are responsible for moving the penalty median. 00:01:05 but pre-flood they wouldn't pay higher fee due to lack of backlog? 00:01:39 but i'm not sure, no. it's just what i have read from others. higher fees -> faster block size increase. 00:01:53 Sure, but I am saying it doesn't matter how much a steady 100 kB pays in fees if there are never any high fee txs available when it is time to weigh fees against the penalty. 00:02:38 The miners run out of high fees before they get to the penalty median, and are left with the same choice: Do I include another minimum fee transaction, and pay this penalty? 00:03:05 maybe the entity will update to 18.3.2 :D 00:03:22 In other words, you need enough higher fee txs to consistently fill blocks to the penalty median for them to matter for block size expansion. 00:04:46 preland: I wrote a simulation of the scaling algorithm. If you give me a tx volume I can run the simulation with minimum fees and see what happens. 00:07:30 I can give you different values, but the tx volume has changed a bit over time so it is hard to say what will actually happen. 00:11:01 Doing lazy math, let's say we are seeing 2tx per second. That is 240 tx in a block, with each flood tx being ~1.5kB. So we need 360 kB blocks to consistently clear that. The network should get to that about a day from now. 00:11:36 In reality I think we are seeing less than that at the moment, but volume has been variable and it wouldn't surprise me if it increased further. 00:11:47 <123bob123:matrix.org> Is that out? 00:12:49 might have already started using it ;) 00:13:33 /s 00:13:38 i think we should talk about how to adopt the situation; nodes being slow and poisoned outputs. 00:14:24 no you get ridiculed for asking questions, its all organic 🤣 00:15:13 lets assume its organic, still nodes being slow which is problem. 00:15:34 free benchmark on mainnet. 00:15:38 i am sure devs are looking into it 00:16:10 decoy selection, is also important 00:16:43 plus fee algo, which chain makes it cheaper to spam as the scale increases 00:17:10 The fees only reduce after the long term median adjusts. That happens ~70 days from now. 00:18:24 we have improved node and wallet performance in v0.18.3.1 – but a lot of people haven't upgraded which causes extra load on nodes 00:20:24 Well, if happens 70 days from now if volume is consistently above 300 kB for the entirety of that time. 00:22:45 Hey I'm gonna post the XMR node on the SimplifiedPrivacy server. Should we get a list going of semi-public nodes that work? 00:24:19 <123bob123:matrix.org> There in feather and monero gui already 00:25:00 You mean the node list is updated? 00:25:02 15.204.197.8:18081 00:25:31 <123bob123:matrix.org> 192.168.1.150:18081 00:25:51 10.150.0.22:18081 00:25:53 111.90.140.158:18081 00:25:55 https://matrix.monero.social/_matrix/media/v1/download/monero.social/QMJorJXvfOKBsGJVAcFlHZja 00:26:02 BTW, I think one of the attack gone or something 00:26:14 they failed to bring my node down 😂 00:26:47 RPC latency went down an order of magnitude 00:26:53 <123bob123:matrix.org> Noob disable hibernate in power options 00:46:06 Lol 00:49:57 is the congestion a serious problem? 00:51:27 Congestion was caused by a now fixed bugs that prevented the auto fee selection algo to choose higher fees. 00:51:27 It's fixed now, and all wallet will pick up the fix real soon 00:51:48 That fix will also help the blocks to grow faster 00:53:59 maybe time to try "start_mining_for_rpc" feature? 00:54:53 awesome!!!!! 00:55:23 will a similar issue happen again 00:56:59 @ dan yes bitcart uses seths by default. Its hardcoded. But you can change it 00:57:28 "Hey I'm gonna post the XMR node on the SimplifiedPrivacy server. Should we get a list going of semi-public nodes that work" no 00:58:00 "maybe time to try "start_mining_for_rpc" feature?" Just force spammers to pay the actual cost of the spam - 5x or greater 00:58:53 The custom fees theyre using arent benificial either. I see no reason to allow fees outside of the calculation. Shouldnt be able to send a 200001 picp/byte tx 00:59:01 20001* 01:01:36 "It's fixed now, and all wallet will pick up the fix real soon" < half of a fix, as it improves UX for users, but also allows for a cheaper attack 01:02:44 And the big issue with the "attack", now, seems to be that its effecting the ability for nodes to stay in sync of they are being used for services 01:02:58 ? 01:03:09 which ? 01:03:16 Yes, I've read some wallets unusable 01:03:23 If using public nodes 01:03:34 this is a big problem IMO 01:03:48 Even tor nodes? 01:04:07 Are there ways to make nodes more resistant? 01:04:07 its the nodes. Yeah. I think these nodes should pay attn to rpc traffic and see if they are being used to broadcast the spam 01:04:14 We might have found the culprit 😂 01:04:21 Yeah, dont allow public rpc 01:04:25 I mean probably many nodes just wrong specs 01:04:31 what was it? 01:04:44 Rust monero node 01:04:53 No way 🦀 01:05:10 Yes 01:05:14 will a similar event happen again 01:05:46 Ahhaaha I misread the message 01:06:35 Can't users find the nodes which submit this? Or is this all by remote nodes 01:06:36 Attack seam to have stopped when the words "Tencent cloud computing" where said in another Monero channel 01:06:36 Or one of the attack stopped (the load of my node went way down in the same minute) 01:06:46 i think it may be possible that public rpc nodes are actively being spammed 01:06:57 or does d++ protect too much against any kind of analysis 01:07:14 Yes, node providers should post some data 01:07:18 Ie, spammer isnt using their own node, but trying to take down cakes, feathers, seen nodea 01:07:20 So it was a good idea to not spam nostr 01:07:54 Node providers dont keep logs when possible 01:08:01 Its good that we now found some problems 01:08:27 If this spam continues, is this the longest attack ever on xmr? 01:08:39 No 01:08:52 Yep, they began to hammer my node the second it was added to feather (I initially tought it was feather user but no, the load went byebye but the RPC hit per interval just went down "a little" 01:08:52 1/16 was longer, i think? Maybe im wrong 01:08:54 Define attack 01:08:59 I'm surprised that there isn't like a "monero is hacked" sentiment already pushing price down 01:09:11 And xmr has had oscillating daily traffic for years 01:09:15 Thats the longest i think 01:09:21 Active wide spread network damaging attack 01:09:26 is it possible that this happens again? What caused it in the first place? 01:09:37 Hm. Probably banning ofrnxmr 01:09:42 Yolo 01:09:43 It’s probably because the transaction times are still lower than others 01:09:57 Lol 01:10:19 I wont accept 600k from irs. I need at least 60m 01:10:24 "Tencent cloud computing" caused it, my guess considering they stop when we mentioned it. 01:10:43 huh 01:10:46 Would be cool if we at least got larger ring size until seraphis & fcmp 01:10:51 Fmcp* 01:10:53 someone said the issue was fixed 01:11:03 Dont need bigger rings 01:11:06 The Chinese Communist Party 01:11:07 Need clean ones 01:11:29 Wownero has 22 rings, rheir effective ring size like like 1 01:11:39 the fix in monero code is to prevent legitimate monero transaction to be delayed when such attack occure 01:12:01 Thats one of moneros biggest roflmfao wtf moments 01:12:13 We just keep raising ring size pretty much arbitrarily 01:12:43 The sad thing is, I’m pretty sure someone could do this attack with a small amount of resources 01:12:44 Heck; if I was bored/angry at Monero for some reason and had extra time to burn, I could probably write a modified monero-cli that literally just constantly splits outputs until the wallet is drained 01:12:44 Divide 1 XMR by 20 nano-Nero, and that’s how many bytes I’d shove into mem-pool temporarily, and permanently onto the blockchainZ 01:12:46 Lets not, like, yknow, not add bs to rings. Lets just allocate more space for bs 01:12:47 And we just didn't be aware of the issue because there was no test of this before where the bug would've shown itself 01:12:55 can't we simulate attacks on stage/test net 01:13:00 And look 01:13:02 temporary solution: blacklist last week's outputs 01:13:03 so is it a serious issue or no? 01:13:11 Lol 01:13:16 These attacks have been done 01:13:21 And practiced 01:13:28 Sauron has only 19 rings. 01:13:37 Nah 01:13:59 but only needs one 01:14:41 serious issue? meh. Ux for users is fine 01:15:45 Issues are fixable. Like the rpc race, or the min fee 01:16:04 Better that it happened under an attack than real adoption 01:16:26 Attacks have limited resources. Adoption is forever. 01:16:37 attacks help products mature 01:16:54 True 01:16:54 And the best part? Delist, spam, do whatever you want. were not going anywhere, not even down in price 01:16:54 Actually it’s pretty good this happened before Haveno tried to launch 01:16:54 It already had enough going against it, it doesn’t need anything else to go wrong lol 01:17:29 ofrnxmr[m]: I guess I mean is there something fundamentally wrong with monero? 01:17:54 Right this second? Just bugs 01:18:00 <123bob123:matrix.org> Free pentesting 01:18:07 Fundamentally there is something wrong with everything 01:18:07 Nothing in this world is perfect 01:18:19 But also right this second people are torture testing to find out how to reproduce in a vacuum and how to solve 01:18:28 <123bob123:matrix.org> Not me 01:18:28 Ofrnxmr is perfect. Blasphemy 01:18:40 <123bob123:matrix.org> I am perfect too mum 01:18:44 ofrnxmr* 01:18:50 Lol 01:18:59 My mom told me i was just perfect. Meaning, im not anything else! 01:19:50 But no, at this point in time, unless someone give me 60m, im going bk to sleep 01:19:57 I’m imperfect; that’s why I make the mistakes that others won’t make 01:19:58 I’ll be the biggest idiot in the room so that the one time I actually do something useful it’ll be worth it 01:20:00 Im not worried about these amateurs 01:20:23 Preland.. i am the biggest idiot in the room. Why do you think im so noisy 01:20:32 Instructions unclear: do you accept 50m ElonMoonShotCoin? 01:20:39 no 01:20:43 Xmr only 01:20:51 Ig I can’t argue with that 01:21:17 I think you are going to have to wait till 24,097 to have 60m XMR 01:21:24 Hard to argue with dumb ppl. I would know. I talk to myself more than anyone 01:21:48 400k xmr 01:22:01 Kucoin claimed to have 9m 01:22:08 I dont see what the problem is. Are they broke? 01:22:08 Why can’t we make Monero propaganda with quotes like this instead of the generic “track this” or “privacy good” stuff 01:22:26 Lol 01:22:29 😂 01:22:38 I posted my monero address. Been waiting for them to pay for the service. Cheap asses 01:22:56 I'm a little drunk I might be repeating myself but is this something to worry about or is it sorted? 01:23:10 Tencent is jealous; they want us to use Clash of Clans gems for transactions instead 01:23:29 Partially sorted, the other half is being investigated intensely by a few devs 01:24:00 Most likely no 01:24:01 You won’t be able to buy a coffee with XMR in its current state 01:24:01 But I don’t think your barista would rly want to wait 20m for confirmation time anyways 01:24:06 Other half being the nodes 01:24:21 I can buy coffe with xmr rn 01:24:24 Yeah the node part is what concerns me 01:24:30 I jsut need to use a wallet that has the fee repaired 01:24:46 see.. the good thing? 01:24:48 Hmm 01:24:54 If public nodes all turn ofd their rpc, were find 01:24:57 Fine* 01:25:07 Why wouldn't you be able to buy a coffee with XMR right now? 01:25:23 Users just need to have rpc nodes that they TRUST and stop using default attack points 01:25:30 Why would the barista wait 20m? the tx should be in a block in 2 01:25:30 the monero node was never designed to have tons of connections, something like monero-lws would be better suited for that 01:25:32 If it's tencent, they made a mistake... hahaha 01:25:39 Transaction delays 01:25:39 Although tbh just pay more for fees, it should be fine 01:25:41 https://matrix.monero.social/_matrix/media/v1/download/monero.social/epRQYFbAWnMSTtFPLlJQeHds 01:25:58 Selsta - rpc was clearly not designed to _handle_ a lot of connection 01:26:12 But next to btc, it was clearly intended to 01:26:29 yes monerod wallet rpc wasn't designed to have tons of wallets connected 01:26:57 Cuz the barista is paranoid? Idk, I was just pointing out that the output locking may still be an issue for some use cases 01:26:58 Especially if Monero gets back ported to pre-existing financial systems (no one wants to get “Abagnaled”) 01:27:20 if the tx pool has less than 300kb of 5x fee tx in it 01:27:33 You can be sure youll make it into next block 01:28:00 If whole pool had 5x fees, those tx will clear and blocks will grow much more quickly 01:28:50 Weird question: do the core Monero executables (daemon, rpc etc.) use multi-threading? 01:29:04 why is this 01:29:06 Yes 01:29:25 Ok good 01:29:36 Which one 01:29:44 just use normal fee, or wait for the weekend to pass then update your wallet and it will do it for you 01:30:56 Yeah; although in the future I could see ultra-light UX-focused wallets just wouldn’t have the fee option user-changeable 01:31:27 Imo theres no reason to allow anyone to set any fee lower than auto 01:31:29 Not an issue now, but it wouldn’t be shocking 01:31:30 Plus you can’t change the fee afterwards so your output is locked…..i think? 01:31:31 That's the idea of Auto (for example feather wallet don't have the fee option) 01:31:31 But Auto is what was broken 01:31:37 And no reason to allow anyone to set any fee to any custom value 01:32:27 And that is what is kinda terrifying lol 01:32:42 At least we figured this out now 01:33:13 well, thats why were v0.18 and not v1.18 01:33:50 is monero fucked? 01:33:56 Were still early as far as adoption goes 01:34:06 For now no 01:34:22 maybe if someone competent was paid to do it 01:34:36 This spammer spams as if they get paid by the hour 01:34:46 Ripping off their financiers 01:35:36 monerochan still a virgin 01:38:55 How is auto fee calculated? Is it network consensus? 01:39:10 Or just median or smt got from node 01:39:17 Jeez, we finally get to see the system in action under some serious load and people start freaking out left and right 01:39:35 Yeah IRS great 01:39:36 Its* 01:39:45 Just asking 01:40:53 Oh I was referring to the messages above 01:41:01 1 sec 01:41:18 auto fee is not consensus 01:41:31 auto fee is impossible as consensus 01:41:45 Something like if last n blocks are greater than 80% full, wallets auto bump. Theres a couple rules for wallets 01:41:46 I haz node, I haz wallet = tx with no problem 01:42:22 I don't use your node 01:42:41 I think impossible is a stretch. We change fees at hardforks with a grace period, could do the same in reverse 01:43:19 impossible if it's based on the txpool state 01:43:52 yeah. But wallets also use 80% full blocks as a trigger 01:44:14 Everyone ask Santa for your own node 01:44:19 Which would come after a full pool. Wallets can keep sending high fee when pool is full but blocjs are not 01:45:07 Nodes dont need to require normal fee until the second condition (full blocks), and nodes should reverrt before wallets as well 01:46:29 considering Cat has her own wallet imma gunna get her her own node 01:48:14 https://matrix.monero.social/_matrix/media/v1/download/monero.social/fnXpyKMumYWvdKyEtkukRsdo 01:48:17 Beginning of a possible new wave 01:49:05 wdym 01:49:22 number (CPU usage and mempool bytes) go up 01:49:38 number (Bandwidth, CPU usage and mempool bytes) go up 01:49:59 Here is what I will say: if the party involved is using the “output hydra” method, then their ability to flood transactions increases as they do it (ie each output turns into two, which turns into 4, etc. etc.) 01:50:24 Yeah, using 1:2 for the attack 01:50:30 it's what they are doing 01:50:41 We haven’t seen that 01:50:41 Yet 01:50:42 So either we are fine…..or the worst is ahead of us (or at least for whoever accidentally transacts using auto fee) 01:50:55 Oh that was already confirmed 01:50:56 Hmm 01:51:23