1 2015-11-09 00:00:45 <wangchun> Lightsword: okay, got it: [2015-11-09T00:00:16.713501Z] receive version message: /RelayNetworkClient:42/: version 70000, blocks=0, us=[::]:0, peer=26, peeraddr=...
 2 2015-11-09 00:00:54 <wangchun> Lightsword: should be ok now
 3 2015-11-09 00:01:39 <Lightsword> wangchun, which relay server are you connecting to?
 4 2015-11-09 00:02:12 <wangchun> Lightsword: The server in Beijing
 5 2015-11-09 00:02:34 <Lightsword> wangchun, do you run relayclient on all your nodes with bitcoind?
 6 2015-11-09 00:03:10 <wangchun> Lightsword: We have our own block distribution network among our own servers
 7 2015-11-09 00:04:04 <Lightsword> wangchun, do you use a GBT relay or just connect the bitcoind nodes together with addpeer?
 8 2015-11-09 00:04:11 <phantomcircuit> Lightsword, he shouldn't specify the server, it'll find the best one automatically now
 9 2015-11-09 00:04:34 <Lightsword> phantomcircuit, oh, was that pretty recent?
10 2015-11-09 00:04:53 <wangchun> Lightsword: You may watch if the lag getting improved from now on
11 2015-11-09 00:04:55 <phantomcircuit> more than 2 months ago
12 2015-11-09 00:05:09 <phantomcircuit> it connects to them all in a loop and selects the one that had the lowest latency in connecting
13 2015-11-09 00:06:04 <wangchun> Lightsword: No, we pre-broadcast stratum jobs from mining processes to propagation servers
14 2015-11-09 00:06:44 <Lightsword> wangchun, any idea why bw.com is so slow? they are using ckpool right? I know they are at least using ckproxy to some extent due to the connect message.
15 2015-11-09 00:07:35 <wangchun> Lightsword: bw.com does not have a good technical team, I believe
16 2015-11-09 00:08:46 <Lightsword> wangchun, adding these two options to bitcoin.conf make a huge difference for pools reliant on bitcoind’s GBT minrelaytxfee=0.0001 limitfreerelay=0
17 2015-11-09 00:09:06 <wangchun> hmm I cannot believe the relay network was not in effect in the past few months
18 2015-11-09 00:09:11 <Lightsword> I run ckpool and was getting massive GBT delays of up to 10 seconds
19 2015-11-09 00:09:51 <Lightsword> but those two options brought it down to under a second
20 2015-11-09 00:10:20 <wangchun> Because In August we only got one block orphaned, that is 0.1%, without the relay network
21 2015-11-09 00:10:54 <Lightsword> wangchun, you get template updates over stratum from western pools right?
22 2015-11-09 00:12:55 <Lightsword> wangchun, I know I orphaned one of your blocks recently https://blockchain.info/block-height/381905
23 2015-11-09 00:13:04 <wangchun> Lightsword: We watch stratum updates from bitfury, slush, eligius, btcguild, antpool, bw, btcchina, wemineltc, coinotron and ltc1btc
24 2015-11-09 00:13:23 <wangchun> Lightsword: http://www.f2pool.com/bitcoin-blocks last month we orphaned 4
25 2015-11-09 00:14:24 <wangchun> Lightsword: ghash.io blocked the ip addresses of my stratum clients
26 2015-11-09 00:14:26 <Lightsword> wangchun, you should add the ckpool based pools as well
27 2015-11-09 00:14:34 <Lightsword> ghash.io is very slow
28 2015-11-09 00:14:57 <Lightsword> they are at the bottom of my poolbench tool
29 2015-11-09 00:15:25 <Lightsword> you should add pool.bitclubpool.com:3333 stratum.kano.is:3333 and solo.ckpool.org:3333
30 2015-11-09 00:16:14 <Lightsword> wangchun, does btcchina use the same SPV mining system you do?
31 2015-11-09 00:16:52 <wangchun> Lightsword: I don't like the name "SPV mining", if that means mining on stratum updates, in fact everyone is doing that
32 2015-11-09 00:17:14 <wangchun> Lightsword: I believe it is a in-stock feature of lukejr's eloipool
33 2015-11-09 00:17:35 <Lightsword> wangchun, it is not in eloipool, they grab block headers over p2p but only after full validation
34 2015-11-09 00:18:04 <Lightsword> wangchun, eloipool stock is immune to the BIP66 fork issue you had
35 2015-11-09 00:18:12 <Lightsword> since it waits for full validation
36 2015-11-09 00:18:14 <wangchun> Lightsword: No, eloipool doesn't do full validation
37 2015-11-09 00:18:39 <wangchun> it simply add 1 to the current block height and issue a "clearmerkleroot"
38 2015-11-09 00:18:40 <Lightsword> wangchun, I’m going off of what I was told by Luke-Jr
39 2015-11-09 00:19:16 <Lightsword> wangchun, the way he told me it works is that it waits for the notify before grabbing the block header over the p2p network to bypass GBT
40 2015-11-09 00:20:10 <Lightsword> wangchun, I’m pretty sure only the chinese pools do stratum based SPV mining, and bw.com does not from the looks of it
41 2015-11-09 00:20:12 <Luke-Jr> wangchun: Eloipool doesn't do that until the previous block is fully checked
42 2015-11-09 00:21:02 <wangchun> Luke-Jr: you changed it? When I cloned it in 2013, it was not like that.
43 2015-11-09 00:21:09 <Luke-Jr> wangchun: that's the most it has ever done
44 2015-11-09 00:22:49 <Lightsword> wangchun, were you the first to develop the stratum based SPV mining or do you know of anyone else to do it?
45 2015-11-09 00:22:59 <Lightsword> before you
46 2015-11-09 00:23:03 <wangchun> 72     def updateBlock(self, newBlock, height=None, bits=None, _HBH=None):
47 2015-11-09 00:23:03 <wangchun> If height=None, the new height is set to current height + 1
48 2015-11-09 00:23:21 <wangchun> Again, I don't like the name "SPV mining"
49 2015-11-09 00:23:28 <Luke-Jr> "SPV mining" is a bad name, agreed
50 2015-11-09 00:23:33 <Lightsword> wangchun, what do you prefer?
51 2015-11-09 00:23:40 <wangchun> We developed mining upon other pools' stratum header
52 2015-11-09 00:23:43 <Luke-Jr> wangchun: height+1 is not a problem, since it's obviously always right
53 2015-11-09 00:24:42 <wangchun> but eloipool, at least in my 2013 fork, have a feature to mine on bitcoin p2p block inv command, which does not even have bits and height params
54 2015-11-09 00:24:53 <Lightsword> wangchun, what sort of sanity checks are you able to do to prevent bad data?
55 2015-11-09 00:25:12 <Luke-Jr> wangchun: yes, that inv never gets sent until it is checked