Logged Conversation (all times UTC)
[00:10] <stellar-slack> epsilon: yeah, we're seeing more of that sort of thing than we'd like. some of it is just attempted connections to peers that are offline, but it seems like there are some transport bugs in the overlay layer too
[00:12] <stellar-slack> Appreciate the slices info @graydon, thanks. Is it accurate to say that the current implementation is somewhat UNL-like in supporting only a single list?
[00:12] <stellar-slack> a single list ... where?
[00:14] <stellar-slack> each peer can configure their own peer list and threshold. so long as the transitive closure of the slices chosen by the peers enjoys quorum intersection, the protocol should work. the peers don't all need to choose the same slices or thresholds. we're just not presently checking/enforcing intersection, so it's possible to misconfigure a group of peers without noticing it.
[00:20] <stellar-slack> graydon: thanks!
[00:20] <stellar-slack> I was under the (mistaken, I think) impression that individual nodes would configure multiple slices, as opposed to QUOROM_SET config (a single list).
[00:22] <stellar-slack> hypostatization: well, a list + a threshold is a way of defining a very simple set of slices: the set of all subsets of the list of size > threshold.
[00:22] <stellar-slack> hypostatization: the bug I pointed out is about making a compact encoding of a more-complex set of slices w/o sacrificing notational simplicity
[00:23] <stellar-slack> now that makes sense, thanks
[00:23] <stellar-slack> I mean in theory we could provide the user the ability to define all sorts of sets, but in practice there's a sort of diminishing returns effect as the notation gets more complex. we need to find the right balance of expressiveness vs. compactness.
[00:24] <stellar-slack> at the moment it's a bit too compact / simple, imo. it works but it doesn't handle some of the slices we expect people will want to encode.
[00:24] <stellar-slack> (er, size >= threshold)
[00:26] <stellar-slack> Is the tiered quorum structure in Figure 3 (page 7) possible in the current implementation?
[00:35] <stellar-slack> My understanding is: no, but it would be enabled by #354
[00:36] <stellar-slack> scott: re torquebox rake, have you asked the torquebox devs yet? Seems like they must have some plan there
[00:38] <stellar-slack> hypostatization: I think so. I believe if you make 4 nodes v1...v4 who all have their QUORUM_SET=[v1,...,v4] and THRESHOLD=3; then make v5,...,v8 have QUORUM_SET=[v1,...,v4] and THRESHOLD=2; then make v9, v10 have QUORUM_SET=[v5,...,v8] and THRESHOLD=2
[00:40] <stellar-slack> So it would be achieved via multiple instances?
[00:41] <stellar-slack> I'm not sure I understand the question
[00:41] <stellar-slack> it's a graph of multiple instances
[00:41] <stellar-slack> each vN in that graph is a separate node
[00:42] <stellar-slack> i.e. separate process running on its own machine, exchanging messages, etc.
[00:42] <stellar-slack> got it
[01:00] <stellar-slack> graydon: do you blokes have a UK node up?
[01:01] <stellar-slack> no, we're still just testing in small groups
[01:02] <stellar-slack> I rm everything and restarted - got a UK node immediately :simple_smile:
[01:02] <stellar-slack> from Aussie
[01:15] <stellar-slack> there may well be someone testing the network from the UK presently
[01:15] <stellar-slack> the p2p code connects to as many peers as you configure it to try for, and they flood peer records around
[01:16] <stellar-slack> yeah, I've only ever got you blokes at amazon
[01:16] <stellar-slack> nice to see someone else on the list :simple_smile:
[02:27] <stellar-slack> who is netops?
[02:30] <stellar-slack> for us? matschaffer is kinda in charge of it, but scott, andrew and jed are often quite involved.
[02:39] <stellar-slack> graydon: thanks
[04:02] <stellar-slack> `"age" : 11366`
[04:02] <stellar-slack> `"state" : "Booting"`
[04:02] <stellar-slack> `> recv: 5 before hello`
[04:17] <stellar-slack> epsilon: interesting, what rev is that? Looks like an older state string
[04:18] <stellar-slack> or maybe it went back
[04:20] <stellar-slack> my local vm was one merge back. Trying the latest now
[04:25] <stellar-slack> that VM was one that was swimming along swimmingly like something good at swimming immersed in a fluid that had the viscosity of something exceptionally fluid.
[04:25] <stellar-slack> OT: I kinda wanted to put a u in viscosity.
[04:27] <stellar-slack> matschaffer: a pull and rebuild, you suggest?
[04:29] <stellar-slack> so an older one (run it with --version to see the sha btw). I believe that version was a little hit-or-miss on whether it would sync. It was working for some of us not working for others.
[04:29] <stellar-slack> anyway, the latest would at least get you all the latest fixes, but you may want to keep the previous binary around just incase you want to try that again
[04:29] <stellar-slack> ahh.. it worked for me then. A coupla times.
[04:30] <stellar-slack> I'm building https://github.com/stellar/vagrant-examples/tree/master/single-peer right now, will let you know what it does if anything
[04:30] <stellar-slack> guys I would wait a few days to try to get the nodes connected. It is just going to be a frustrating experience right now. There are some known issues we have to iron out.
[04:31] <stellar-slack> jed: frustrating is okay
[04:36] <stellar-slack> Early daze. I'm amazed with the shit you blokes have pumped out on short notice. I'm sure any reasonable person would expect some bumps here and there.
[04:36] <stellar-slack> matschaffer: thanks!
[04:36] <stellar-slack> cool just warning you :)
[04:40] <stellar-slack> @epsilon: not great:
[04:40] <stellar-slack> ```
[04:40] <stellar-slack> ❯ vagrant ssh -c 'curl http://localhost:39132/info'
[04:40] <stellar-slack> {
[04:40] <stellar-slack> "info" : {
[04:40] <stellar-slack> "ledger" : {
[04:40] <stellar-slack> "age" : 473,
[04:40] <stellar-slack> "closeTime" : 0,
[04:40] <stellar-slack> "hash" : "43cf4db3741a7d6c2322e7b646320ce9d7b099a0b3501734dcf70e74a8a4e637",
[04:40] <stellar-slack> "num" : 2
[04:40] <stellar-slack> },
[04:40] <stellar-slack> "numPeers" : 4,
[04:40] <stellar-slack> "state" : "APP_ACQUIRING_CONSENSUS_STATE"
[04:40] <stellar-slack> }
[04:40] <stellar-slack> }
[04:40] <stellar-slack> ```
[04:41] <stellar-slack> acquiring is spelt correctly
[04:42] <stellar-slack> or.. is is spelled correctly?
[04:42] <stellar-slack> we need a quorum
[04:42] <stellar-slack> google thinks so, they're in my preferred peers list
[04:43] <stellar-slack> ha mat
[04:43] <stellar-slack> :stuck_out_tongue:
[04:43] <stellar-slack> ha! We need the PHD in here NOW
[04:43] <stellar-slack> there is a thing
[04:43] <stellar-slack> nothing especially interesting in the logs for this vm either. Sorta looks just stuck
[04:43] <stellar-slack> yeah, I got that after a reset
[04:44] <stellar-slack> it's got peers anyway, and a lot of `2015-04-14T04:33:19.822 57e320 [Overlay] ERROR readHeaderHandler error: End of file [TCPPeer.cpp:291]`
[04:44] <stellar-slack> It wasn't broke - I should have fucked with it.
[04:44] <stellar-slack> yeah tcppeer too here
[04:44] <stellar-slack> well, it might have very well been broke in some other way :simple_smile:
[04:45] <stellar-slack> it's probably time for me to start trying out https://github.com/stellar/stellar-core/blob/master/test/testnet/run-test.sh as well
[04:45] <stellar-slack> see if I can replicate behavior there
[04:45] <stellar-slack> well, when it connected it stayed connected
[04:45] <stellar-slack> after a reset it baulked
[04:46] <stellar-slack> I was catching other continents 'tho. 1 UK node ftw
[04:47] <stellar-slack> I kinda thought it mighta been donch slumming it at a Manc cafe on his mobile.
[04:48] <stellar-slack> ..and wouldn't that be cool
[04:49] <stellar-slack> who's donch
[04:49] <stellar-slack> donovan
[04:49] <stellar-slack> c something h something
[04:49] <stellar-slack> ah got it
[04:49] <stellar-slack> or something
[04:50] <stellar-slack> fucker's a wizard
[04:50] <stellar-slack> no doubt
[04:50] <stellar-slack> I mean.. sry. He is adept at shit.
[04:51] <stellar-slack> I have to watch my Australianisms :)
[04:51] <stellar-slack> yeah im pretty sure we've got a section on australianisms in the code
[04:51] <stellar-slack> (but yer CEO say frikken)
[04:52] <stellar-slack> no criminals allowed, we'll make an exception for you
[04:52] <stellar-slack> ha!
[04:52] <stellar-slack> all those English procks then
[04:52] <stellar-slack> well, that's you out, Donch
[04:52] <stellar-slack> haha
[04:53] <stellar-slack> bloody white people
[04:53] <stellar-slack> hey btw, survey, have you fired up Horizon API server by chance?
[04:53] <stellar-slack> yeah once
[04:53] <stellar-slack> I don't think it was "doing it" when I tried.
[04:53] <stellar-slack> when did i say frikken?!?!?
[04:54] <stellar-slack> potty mouth!
[04:54] <stellar-slack> oh, ok, guilty as charged
[04:54] <stellar-slack> there's my fav ceo in the world
[04:54] <stellar-slack> :speak_no_evil:
[04:54] <stellar-slack> :hear_no_evil:
[04:55] <stellar-slack> :see_no_evil:
[04:55] <stellar-slack> @epsilon: did you build anything to stellar/ripple before?
[04:55] <stellar-slack> yeah
[04:56] <stellar-slack> nice, client side stuff or what did you build?
[04:56] <stellar-slack> I mean no, I just hang out here because you cunts are awesome, or something.
[04:56] <stellar-slack> 'straliacunt?
[04:57] <stellar-slack> ^ there ya go
[04:57] <stellar-slack> ausfag
[04:57] <stellar-slack> reason i ask is, we've got a pretty brand spankin new API
[04:57] <stellar-slack> and general way of doing transactions
[04:57] <stellar-slack> yes, yes you do
[04:57] <stellar-slack> wanted to get your feedback on it
[04:57] <stellar-slack> transaction envelope, operations, etc
[04:58] <stellar-slack> you know with feedabck comes "I'm not commenting on the shit that was done correctly"
[04:58] <stellar-slack> because that was implied
[04:58] <stellar-slack> I only want to bitch and moan
[04:59] <stellar-slack> because "freedom!"
[05:00] <stellar-slack> I know, I know, I need to throw a telephone at a Jew an then paint half my face.
[05:00] <stellar-slack> and type in all caps...
[05:00] <stellar-slack> im not gonna even attempt to understand that
[05:01] <stellar-slack> while making a fkn movie about Jesus
[05:01] <stellar-slack> yer not Aussie.
[05:01] <stellar-slack> ya, lost me there. aussie speak
[05:01] <stellar-slack> (it's thick(
[05:01] <stellar-slack> no biggie, mate.
[05:01] <stellar-slack> fosters
[05:01] <stellar-slack> blah
[05:01] <stellar-slack> sometimes ya get lost in the weeds
[05:01] <stellar-slack> @joyce coopers, yo
[05:02] <stellar-slack> one day yer kidzz will look back and say "Dad? wtf? It was perfectly contiguous!"
[05:02] <stellar-slack> and then you google
[05:03] <stellar-slack> too late, dad
[05:03] <stellar-slack> kidzz pwn!
[05:05] <stellar-slack> I apologise for treating /dev like a talkshow
[05:05] <stellar-slack> we need an oz channel
[05:05] <stellar-slack> but if Mel was here.. Look Fucken Out.
[05:05] <stellar-slack> omg mel
[05:05] <stellar-slack> it all just came together
[05:06] <stellar-slack> ha!
[05:06] <stellar-slack> I love you guys. You took Rupert
[05:06] <stellar-slack> he lived just down the road..
[05:06] <stellar-slack> how's TV for you guys?
[05:06] <stellar-slack> yeah.. fucked.
[05:07] <stellar-slack> yo @epsilon lets move this conversation to
[05:07] <stellar-slack> #aussie-speak
[05:07] <stellar-slack> yes
[05:08] <stellar-slack> See, that's why you;re the boss.
[13:07] <stellar-slack> question: are secret keys going to be the same for the new and old protocols? As in, are we going to have a smooth transfer or how will this work?
[14:45] <stellar-slack> kwelstr: I believe the transfer from the legacy system may require making a new account.
[14:52] <stellar-slack> @kwelstr: secret keys (the secret seed, in ed25519-speak) will stay the same. The only reason account addresses will change is because we’re no longer running the public key through `SHA256 * RIPEMD160` before encoding to base58check. I’ve been playing with a scheme to help smooth the transition to new addresses by extending the federation protocol, but don’t have much progress yet.
[15:07] <stellar-slack> Thank you Scott and Epsilon
[15:20] <stellar-slack> np :simple_smile:
[18:52] <stellar-slack> request for review: https://github.com/stellar/horizon/pull/17
[21:05] <stellar-slack> here is one for the interested viewer - now that things are refactored, it's simpler to figure out when certain things are not right in herder https://github.com/stellar/stellar-core/pull/442
[21:05] <stellar-slack> damn it, internet explorer. http://caniuse.com/#feat=eventsource no support for SSE, even after this long
[21:05] <stellar-slack> luckily it can be polyfilled
[21:53] <stellar-slack> rfr: https://github.com/stellar/horizon/issues/18
[21:54] <stellar-slack> Going to build out a prototype of ^^^ now
[23:36] <stellar-slack> @monsieurnicolas: are the new changes worth rolling out to testnet or should I wait for more?
[23:37] <stellar-slack> yes, with the caveat that catchup is still waiting on Graydon's change to be reliable
[23:37] <stellar-slack> k
[23:37] <stellar-slack> should have something soon
[23:39] <stellar-slack> @scott: Do you know about the `Last-Event-Id` header for with SSE?
[23:39] <stellar-slack> http://www.w3.org/TR/2011/WD-eventsource-20110310/#processing-model
[23:39] <stellar-slack> It's crying out to be used for the "last seen ledger" :simple_smile:
[23:56] <stellar-slack> I'm wiping the testnet
About StellarVerse IRC Logger
StellarValue IRC Logger
is part of