Hide Idle (>14 d.) Chans
← 2022-09-17 | 2022-09-19 → ↓
dulapbot: Logged on 2022-09-17 21:42:25 mats: assigning blame to the wrong party here -- it is the government propping up this scheme, not bankers
asciilifeform: and not simple matter of 'they ordered' -- reich cannot operate w/out printolade; printer cannot operate w/out the 'kinetic' component of fiatola hegemony
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-09-17#1113834 << pretty close to the '% who eat recycled food' neh
dulapbot: Logged on 2022-09-17 21:53:57 mats: in other news, TIL 73.6% of US adults over 20 are overweight or obese https://www.cdc.gov/nchs/fastats/obesity-overweight.htm
asciilifeform: ( lulzily in ru to this day most common reason for rejecting a conscript is 'underweight' )
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-09-17#1113837 << ru lost these colonies in sovok collapse, so they were picked up (originally by usg, nao cn will)
asciilifeform: mats: orthogonally, are you interested in getting on pestnet?
asciilifeform: other folx would prolly comment interestingly re e.g. the noose linx mats digs up
phf: awt, i seem to have lost peering. is your ip still 200.122.181.26?
phf: actually not sure what happened but none of the peers on my list (ascii, trinque) respond
phf: also i tried changing protocol versions, still nothing
dulapbot: (pest) 2022-09-14 asciilifeform: PeterL: 100.15.116.69:1337
asciilifeform rolled back to blatta 9973, which dun have addrcast yet
asciilifeform: phf: does your cl pestron have it ?
phf: to quote some 90s comedy "i ain't seeing shit"
asciilifeform: phf: from all peers or only awt & asciilifeform ?
phf: that's the only two peers i've ever really tested, so i'm not sure if other creds are still even legit
asciilifeform: should be, if folx threw the keys in their stations, blatta dun have rekey yet
phf: but i sent a couple of broadcast and prod messages to the ip you gave, with version FB, and i'm not getting any kind of reaction
asciilifeform: phf: is yours in a nat ?
phf: yes, but it's always been in a nat
phf: i open up one socket, which i reuse for all things, and as soon as the socket is up, there's a 10s garbage packet going out to keep the outbound/inbound alive
phf: but nothing has changed to this strategy
asciilifeform: phf: in asciilifeform's at yer currently 137.103.133.228:16928 ( 17 july ) , which loox like an ephemeral port, it prolly closed
asciilifeform: doesn't explain why not connects w/ asciilifeform's tho (oughta, asciilifeform put in a fwd rule)
phf: the ip has also rotated, but that means that you're not getting anything from me, which should be at this point a mixture of garbage, broadcast and prod
phf: what's the pest version number that your blatta.py says?
phf: perhaps my upstream decided to filter suspicious udp packets
phf: no, pest protocol
asciilifeform: ( lives in message.py )
phf: well, that didn't make any difference
phf: would you kindly mind posting your at table to a paste
phf: i suspect you have fresh awt ip there also
phf: when i finish my pest™ it'll have a web accessible at table, for bootstrapping
asciilifeform: phf: try peering w/ dulapbot btw. key & addr.
phf: asciilifeform: is 9973 the previous stable version? i.e. whatever you guys were running for weeks before?
asciilifeform: phf: 9973 is current 'stable' blatta
asciilifeform: on dulapbot still running 9976
phf: right
asciilifeform: phf: btw try port 38144 for asciilifeform's station ( loox like from dulapbot's pov he's on an ephemeral? and fwd rule is borked ? )
asciilifeform: web config (or otherwise outtaband) at would be quite handy
asciilifeform: dulapbot ftr dun have nat issue, it is in the rack, so oughta work
phf: asciilifeform: yeah, i'm going to try dulapbot instead but need to get to machine with keys. i don't think sending to your ephemeral port, without you also sending to my ephemeral port will work. because and since we're both behind nat, we have to warm up connection state tables
asciilifeform: phf: appeared to work w/ PeterL on 9973 (where addrcast), whether worx depends on what kinda nat (typically ignores origin, but some do not)
asciilifeform: we'll need this, eventually, asciilifeform suspects, for 'final solution to nat'
dulapbot: (pest) 2022-09-07 asciilifeform: afaik with 'symmetric' nats of the type apparently victimizing jonsykkel , the only drill that worx is to hammer random ports, from both directions, until match
asciilifeform must bbl
asciilifeform: err, 9972 where addrcast
asciilifeform genuinely bbl
phf: asciilifeform: addrcast as far as i understand by design (it's your design!) uses intermediary to tell both stations what corresponding epthemeral ports are. so that station start broadcast (of e.g. garbage) and there's a few packets lost, before mutual state table relationship is established
phf: so e.g. a, ip 1, port 1, b ip 2, port 1: a sends to b, now a's router knows that there's 1:1 - 2:1 connection, but this packet is rejected by b, until b ALSO sends something to a, and now b's router knows 2:1 - 1:1 connection. b's packet will go through, if it went after a's packet. and after b sent his packet, a will start recieving packets also
phf: (aka hole punching technique, but i've explained it in a detail above, because people might forget or not know what hole punching actually entails)
phf: you could probably just send the whole 65536 range, but i like the idea of sitting there, trying to hit random ports. sort of like digging an prison escape tunnel with a robotic spoon
phf: blatta network is not being healthy. so i did a transmit through dulapbot, which appeared on logs.nsl about 2 minutes later, and it's been another 3 minutes and it's still not on bitdash
phf: trying to send packets through awt, with updated AT and i'm not getting any reaciton
asciilifeform: phf: loox like it got logged on bitdash in the end
phf: yeah, 10 minutes later
asciilifeform: phf: there's defo a bug w/ embargo logic in all blattas to date ( partly asciilifeform's fault, that section of spec is laughably incomplete )
asciilifeform: phf: the random dig algo is not only aesthetically appealing but afaik actually faster, statistically, than any possible deterministic one
phf: well, i'm getting packets from awt, but sending packets out seems to take Very Long Time™
asciilifeform: phf: likely because he aint peered w/ dulapbot , and so it walks from the latter to asciilifeform then to him, and each time sits in the hearsay clink
phf: i'm talking to him directly now
asciilifeform: a and still slow?
phf: yeah, i sent "test", got back message type 3, and it's been sitting couple of minutes now (shouldn't that buffer be 10s?)
asciilifeform: ( iirc awt himself is on 9972, where errything dog slow, it gets into prod loops and devours the cpu , was wai asciilifeform rolled back )
phf: but i sent him prod and got it back instantly
asciilifeform: interesting
phf: in fact prod followed by getdata got me back a packet at eval times
phf: but my "test" packet's been in flight for 5 minutes now
phf: and it just landed
asciilifeform: there iit went
← 2022-09-17 | 2022-09-19 → ↑
Random(asciilifeform) | Download hourly DB snapshot | Get Source Code