Hide Idle (>14 d.) Chans
← 2022-01-29 | 2022-01-31 → ↓
asciilifeform: billymg: eggog 500 on pestlogger www ?
billymg: just noticed, looking now
billymg: looks like postgres crashed. trying to start / restart / stop it gives: pg_ctl: PID file "/var/lib/postgresql/10/data/postmaster.pid" does not exist
asciilifeform: billymg: iirc that's the lockfile. start oughta work
billymg: start gives: * WARNING: postgresql-10 has already been started
asciilifeform: kill it manually (if alive), rm the lockfile, oughta start
billymg: but no postgres processes turn up
billymg: `ps aux | grep postgres` no results
billymg: do you know off hand where the lockfile is?
asciilifeform: dmesg dun show any disk eggogs, does it ?
asciilifeform: the pid thing ~is~ the lockfile iirc
asciilifeform: ( check /var/run/postgresql/ also for pg liquishit )
asciilifeform not witnessed this particular barf afaik
billymg: dmesg shows some out of memory errors actually
billymg: Out of memory: Kill process 16688 (postgres) score 186 or sacrifice child
asciilifeform: nuffin suggestive of ssd demise, neh ?
billymg: Killed process 16688 (postgres) total-vm:685136kB, anon-rss:1612kB, file-rss:367580kB
asciilifeform must, unfortunately, to bed
billymg: asciilifeform: can't say i'd know what to look for re: disk problems
asciilifeform: billymg: eggogs featuring /dev/sda, sdb
billymg: nah, none of those
asciilifeform: aite then soft barf
asciilifeform will bbl, will help investigate tomorrow if billymg not finds pill earlier
billymg: gonna restart the box
billymg: well, restart didn't appear to help
billymg: ah, wait, now it's back
billymg: so looking at the postgres logs it looked like it indeed kept getting kill-9'd by the os
billymg: prior to that it was getting some bogus queries through the crawler www, e.g. lookup info where host=drupal.php
billymg: so possibly the site got it by a bot that was trying to sql inject
billymg: none of the queries had any bad effect, other than wasting pg resources (my fault for not having sanitization in place at the app level)
billymg: i backfilled the lines in #a from asciilifeform's logger. only one line was dropped in pest (which i pasted in the channel). i've added 8gb of swap to the rk for now but will have to plug the unsanitized client input hole tomorrow
cgra: http://logs.nosuchlabs.com/log/asciilifeform/2022-01-29#1077198 << note that, peculiarly, the genesis hash was nominally ~2600x harder to mine than the genesis difficulty required! for easy check, see next block hash: ~1300x larger integer than the genesis hash.
dulapbot: Logged on 2022-01-29 16:06:04 whaack: asciilifeform: the genesis starts with a decently high difficulty, it would take a while to generate the synthetic blocks on a home computer
cgra: sometime earlier, i made two experiments to verify some upper limits of blocks. at the time they weren't valid blocks, but i rectified that in the mean time: it took perhaps <1 hour per block to mine, with an 8-core CPU, and a python(!) miner
dulapbot: (therealbitcoin) 2020-10-03 cgra: asciilifeform: re your ada piece, i hand-crafted two versions of #1 block (although with invalid hash and merkle tree, had to temporarily disable those specific checks to experiment).
cgra: asciilifeform: btw, did you ever comment on this one? valid block species now here and here (obv the eating node must've chain height below the builtin checkpoints)
dulapbot: (therealbitcoin) 2020-10-03 cgra: one of them had ~1MB output script, the other had ~112k outputs, and they both got accepted by trb (fresh install, 0% synced). i saw much tighter limits set in your code
asciilifeform: cgra: oh hm, must've missed somehow?
asciilifeform: cgra: how the hell does one stuff ~112k outputs in <= 1e6 bytes tho
asciilifeform rereads nqb structs, can nao see how
asciilifeform when gets a chance to pick ^ up again, will have to properly calculate the maxima...
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-01-30#1077242 << this kinda thing used to happen w/ early phuctor regularly
dulapbot: Logged on 2022-01-30 00:57:59 billymg: none of the queries had any bad effect, other than wasting pg resources (my fault for not having sanitization in place at the app level)
cgra: whaack: i'm possibly not aware of all implications, but i keep contemplating a 'pesttronic block explorer' accepting read-only sql queries from wot
cgra: for example, the last dig i did was "what's the last block mined with the hard-coded, minimum difficulty the genesis block started off with?"
signpost: cgra: yep, imho this can be generalized to perhaps an async RPC mechanism.
signpost: would be quite marvelous to interact with "bots" in this manner, where you pass them a structured request, and receive a structured response back via pest.
signpost: the IRC request -> paste response mechanism is a rough approximation that can be much improved.
signpost: http://logs.nosuchlabs.com/log/asciilifeform/2022-01-17#1073543 << what I meant here by UCI miners was boxen being paid to run jobs distributed over wot.
dulapbot: Logged on 2022-01-17 12:33:59 signpost: (and worthwhile to consider how to allow folks to turn arbitrary boxes into "UCI miners" as easily as possible, i.e. "run this here proggy, contains a whole linux within")
signpost: old thread on subj for those not familiar with the UCI macro http://logs.nosuchlabs.com/log/trilema/2017-07-20#1687384
dulapbot: (trilema) 2017-07-20 asciilifeform: i want a nonstar topology.
thimbronion: speaking of old threads - bitcoin-assets appears to be gone.
asciilifeform: thimbronion: gone from where ?
signpost: but upstack, the logical progression of IRC botworks is being able to (deedbot:deed "signed payload to be deeded") into pest, and receive back an ack that it was heard.
signpost: obviously s-exp is perhaps not the best encoding, just giving it as a placeholder.
signpost: asciilifeform: looks like the www, unless I forgot the url
thimbronion: asciilifeform: the domain http://log.bitcoin-assets.com/ is gone
signpost: *iirc
dulapbot: Logged on 2021-07-06 15:46:47 kakobrekla: i took that shit down recently, i only kept it up so long cause alfie said im gonna fold before mp does or some shit like that
thimbronion: asciilifeform: ah - missed that
asciilifeform: gone for at least that long
signpost: kako was in the end a pretty small-ego'd guy, seems like.
signpost does not have much use for people whose principles flow only from what pays them.
thimbronion occasionally checked to see what they were currently making sarcastic remarks about
signpost: sad when folks make their only basis of relationship bawing about old wounds.
signpost is here because the decentralization tech tree still grows.
signpost: back in a bit.
asciilifeform: thimbronion: iirc by the time kako unplugged that thing, no one had said ~anyffin in yr+
asciilifeform: signpost: wasn't clear to asciilifeform that kako ever had anyffin in the way of principles
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-01-30#1077257 << imho the correct place for a block exploder is ultimately a pseudoserver within trb noad per se, locally
dulapbot: Logged on 2022-01-30 13:00:46 cgra: whaack: i'm possibly not aware of all implications, but i keep contemplating a 'pesttronic block explorer' accepting read-only sql queries from wot
asciilifeform: (can expose the port to net at large, if operator wants, naturally)
dulapbot: Logged on 2022-01-30 06:05:04 cgra: http://logs.nosuchlabs.com/log/asciilifeform/2022-01-29#1077198 << note that, peculiarly, the genesis hash was nominally ~2600x harder to mine than the genesis difficulty required! for easy check, see next block hash: ~1300x larger integer than the genesis hash.
asciilifeform: hash of gen block was 000000000019d6689c085ae165831e934ff763ae46a2a6c172b3f1b60a8ce26f; at diff==1, needed 2 fewer leading 0's than he had
asciilifeform: or hm, iirc needed 32 leading 0 bits at diff1. which'd make the genesis, with 43 leading 0's, 2^11 i.e. '2048x harder than had to'
asciilifeform if slipped sumwhere, invites folx to correct.
asciilifeform: ( 0 ~bits~ that is )
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-01-30#1077262 << fwiw asciilifeform still waiting for sumbody to invent cpu work that's actually worth the overhead (and snoop risk) of farming out to fleets of boxen
dulapbot: Logged on 2022-01-30 13:31:39 signpost: http://logs.nosuchlabs.com/log/asciilifeform/2022-01-17#1073543 << what I meant here by UCI miners was boxen being paid to run jobs distributed over wot.
asciilifeform: ( mining per se was ~briefly~ it, gpu/asicism quickly made such schemes -ev tho, in '13-14 already faded to ~nuffin (supposing was ~ever~ +ev at all)
dulapbot: Logged on 2020-05-15 21:00:56 asciilifeform: lru: w/ bitcoin proper, as late as 2013 there were trojan payloads that (with variable success -- variations in iron/drivers made this nontrivial) gpu mined.
asciilifeform: trb noad operation was sometimes quite erroneously given as example of 'useful' payload for hypothetical bots; but if you think about it, it takes considerable sweat even for ~willing~ operator to maintain a noad worth connecting to
dulapbot: Logged on 2022-01-17 13:12:59 asciilifeform: archaetypical 'botnet' aint actually as useful as folx naively think.
asciilifeform: there'sa reason wai most botnets historically used for brute ddos-for-hire.
asciilifeform: there simply aint much else anyone knows how to do with'em atm.
asciilifeform: ( possibly still used for e.g. premining newly-perpetrated shitcoins? i.e. before the gpu/asic folx cotton on. nfi. )
asciilifeform: ftr on a typical bot box you dun have the cpu all to yerself, either -- often enuff will share it w/ over9000 similar shits
asciilifeform: ( + the luser himself running a 40GB msoffice etc lulz )
asciilifeform: 1 hypothetically interesting application may be 'mining' the mswin rng in conjunction w/ prb privkey gen.
asciilifeform: (would require storing the 'utxo set' tho)
asciilifeform: storing on ea. victim, that is.
asciilifeform: billymg: nifty. not 100% sure why would want to use a dynamically-linked bitcoind tho (i.e. where 'ldd' returns anyffin)
asciilifeform: (ftr on a rotor's built ldd returns nuffin)
asciilifeform: on a musltronic gentoo, one oughta be able to , theoretically, build exactly same thing as rotor's output, but w/out the 'buildroot' liquishit in the process
asciilifeform: (i.e. getting same exe)
billymg: asciilifeform: i almost mentioned that at the end of the article, i recalled the rotor built bins being static
signpost: not botnet, but rather infrastructure for wot members. for example, I would happily trade cycles to remove a single point of failure for asciilifeform's www
signpost: "botnet" is the wrong metaphor, assumes all-comer boxen can be fungible.
signpost has one such application in his head, that for example his l1 may in time be able to purchase a burner laptop while traveling, and call down a usable system into it quickly from wot-net.
signpost: to hazard a sketch, I'd create a travel key, and distribute a disk image among friendlies in l1. after arrival, I'd ping those friendlies for a swarm download of the disk image with travel key, and appear on pestnet.
signpost would rather pay an in-wot person to run a signed job containing his blog and be spared the details of deployment than continue to do sysadmin work.
signpost: there are lots of such use-cases that are applications of decentralization among unequal participants.
signpost: now, if anybody has a *real* breakthrough in homomorphic encryption at >glacial processing speed, changes quite a bit.
signpost: http://logs.nosuchlabs.com/log/asciilifeform/2022-01-30#1077314 << works fine on my item. will get the thing into folks' hands sometime tomorrow, almost done massaging vpatches.
dulapbot: Logged on 2022-01-30 19:21:00 asciilifeform: on a musltronic gentoo, one oughta be able to , theoretically, build exactly same thing as rotor's output, but w/out the 'buildroot' liquishit in the process
signpost: perhaps billymg wants to cut the vpatches for openssl, bdb, boost at that point
signpost: would be great to lop off every turd in all of them not actually *used* by trb, and this kind of work was precisely why I bothered to make pentacle.
signpost: and why taking pains to avoid arbitrary dependencies; the thing oughta pull *inward* from here.
signpost: while on the subj, wouldn't mind feedback on whether to have one single v-tree for all possible system components, or each piece of software having independent lineage.
signpost: I've opted for latter to avoid regrind hell when folks disagree about what they want on their systems, but curious if others disagree.
signpost: in my mind, each system is defined by the patches/seals/wot they have locally, which comprises what *can* be built, and out of that systems are composed with what's chosen to build.
signpost: perhaps later we find there's coordination overhead in not having one canonical src tree, as is done in say bsd ports.
signpost: not insurmountable to merge later.
dulapbot: Logged on 2022-01-30 20:24:14 signpost: not botnet, but rather infrastructure for wot members. for example, I would happily trade cycles to remove a single point of failure for asciilifeform's www
dulapbot: Logged on 2021-11-28 19:37:11 asciilifeform: signpost: somewhat apropos, thought of your lubytron in context of possible pheature: pestron takes a local dir and 'hosts'. peers (and optionally broader pestnet members, e.g. l2/l3) can visit e.g. http://localhost:8000/signpost and see his 'www'.
dulapbot: Logged on 2022-01-30 20:25:22 signpost: has one such application in his head, that for example his l1 may in time be able to purchase a burner laptop while traveling, and call down a usable system into it quickly from wot-net.
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-01-30#1077320 << this'd be a++. tho as asciilifeform understands, ~100% of the problem is 'bake a linux which actually boots on xyz random irons' rather than distributed-anyffin per se
dulapbot: Logged on 2022-01-30 20:27:27 signpost: to hazard a sketch, I'd create a travel key, and distribute a disk image among friendlies in l1. after arrival, I'd ping those friendlies for a swarm download of the disk image with travel key, and appear on pestnet.
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-01-30#1077323 << asciilifeform usedto work in that racket and would bet heavily against it ever at any pt amounting to anyffin other than 'moar studies needed'(tm)(r) 'climate science'
dulapbot: Logged on 2022-01-30 20:32:00 signpost: now, if anybody has a *real* breakthrough in homomorphic encryption at >glacial processing speed, changes quite a bit.
dulapbot: (trilema) 2017-09-15 asciilifeform: kanzure: i spilled the beans from a similar darpa conference that i attended, in the heart of the beast itself, few yrs back ( it's in the l0gz, spoiler : multilinear map homomorphic crypto is bunkum ) and still waiting for gasenwagen
asciilifeform: ( most hilariously, even the impossibility proof of the general case did not put a damper in the flow of moolah. )
dulapbot: (trilema) 2015-07-07 asciilifeform: other pertinent maths include the proof of why 'homomorphic obfuscator' is impossible in the general case
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-01-30#1077329 << imho things which are actually separable, oughta be separate. (may be easier said than done, tho)
dulapbot: Logged on 2022-01-30 20:43:55 signpost: while on the subj, wouldn't mind feedback on whether to have one single v-tree for all possible system components, or each piece of software having independent lineage.
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2022-01-30#1077332 << unless we throw out the manifests thing (doing which has own potential minefields) the coordination overhead of '1 tree' would be quite lethal (e.g. 'hey you gotta regrind yer postgres patch, cuz i changed the default sound card config 5min ago')
dulapbot: Logged on 2022-01-30 20:46:26 signpost: perhaps later we find there's coordination overhead in not having one canonical src tree, as is done in say bsd ports.
← 2022-01-29 | 2022-01-31 → ↑
Random(asciilifeform) | Download hourly DB snapshot | Get Source Code