log☇︎
▂▂▁▁▁▁▁▁▁▃▁▁⏐︎▂▃█▂▂▂▂▁▁▁▁▄ 2479
asciilifeform: soo, in strange findings: on dulap, 'time curl http://logs.nosuchlabs.com/log' yields 2.342s; ( from my chair, same -- 2.455s ); but on identical (3m ago) clone of db , running locally, on substantially slower box, 0.252s !
asciilifeform: dulap aint currently swamped with loads, either
asciilifeform: ( on either log or phuctor )
asciilifeform: i'm beginning to suspect that the indices work here (pg 9) and not there (10)
asciilifeform: i.e. silently fail
asciilifeform: ... which is odd, cuz phuctor's -- work
asciilifeform: interestingly, seems to vary there by factor of ~2x !
asciilifeform: but!! only via apache ! i.e. time curl http://127.0.0.1:5002/log dun show either the variance or the slowdown.
asciilifeform: culprit, apparently , the frontend
asciilifeform: ... 100% of the time diff, apparently, in response req time
asciilifeform: err, nginx there, presently. interestingly, set up as experiment 'wsgi', python's equiv. of php's 'mod-php', and made ~no measurable diff !
asciilifeform: actual generation of e.g. current /log, takes 0.073s !
asciilifeform: approx 0.2s is eaten by transmission on net from BingoBoingostan to my chair; the rest is eaten by the www server somehow
asciilifeform: cachism is apparently red herring, the proggy is actually fast enuff. there is parasitism in the loop.
asciilifeform: reason why never noticed this effect with phuctor, is that its pages are small
asciilifeform: i'ma have to set whole orchestra up on a box w/ classic apache, to see whether same thing
asciilifeform: ( and if cures, move dulap entirely to classic apache. back in the old days went w/ 'nginx' because slightly moar miserly re memory )
asciilifeform squeezed out a bit of the delay via knob-turning but this is still ridiculous
asciilifeform bbl,meat
asciilifeform: nao almost tolerable
feedbot: http://blog.lobbesblog.com/2019/08/znc2tmsr-vpatch/ << lobbesblog -- znc2tmsr vpatch
lobbes: ^ asciilifeform, et al.
lobbes: http://logs.nosuchlabs.com/log/trilema/2019-08-11#1927288 << archiver back online btw (it has been chugging through the backlog for the last 4 hours or so)
snsabot: Logged on 2019-08-11 09:25:32 mp_en_viaje: http://logs.nosuchlabs.com/log/trilema/2019-08-10#1927269 << cool deal.
lobbes bbl sleep
mp_en_viaje: http://logs.nosuchlabs.com/log/trilema/2019-08-12#1928191 << you know, someone fucked something up somewhere, because recall http://btcbase.org/log/2019-07-14#1922400 ? sounds exactly the same, server just sits there for MINUTE before serving anything, for no reason. this didn't use to be the case, i suspect some "upgrade" broke the stack you two both happen to use ? ☝︎
a111: Logged on 2019-07-14 08:12 mp_en_viaje: so since thelastpsychiatrist.com went all fucking stupid (cloudflare aside, and mountain of pointless js aside, there's a spurious 1min delay before any pageload also now!), i scraped the article list both from archive page and category page, and downloaded the list.
snsabot: Logged on 2019-08-12 22:07:05 asciilifeform: i'ma have to set whole orchestra up on a box w/ classic apache, to see whether same thing
mp_en_viaje: http://logs.nosuchlabs.com/log/trilema/2019-08-13#1928198 < tyvm!
snsabot: Logged on 2019-08-13 01:29:02 lobbes: http://logs.nosuchlabs.com/log/trilema/2019-08-11#1927288 << archiver back online btw (it has been chugging through the backlog for the last 4 hours or so)
mp_en_viaje: https://www.wsj.com/articles/new-york-city-considers-lifting-hostel-ban-to-lure-more-tourists-11565649721 << in other fucking lulz, new york joins the chase for norse retards.
mp_en_viaje: maybe they could add the fat black woman at airport to dole out cabs to a queue ? god knows new york's prime for that.
feedbot: http://trilema.com/2019/bring-it-on/ << Trilema -- Bring It On
mp_en_viaje: hoooly shit, i woke up by myself.
asciilifeform: mp_en_viaje: plox to try the 'heavy' log pgs, http://logs.nosuchlabs.com/log/trilema/2019-08-12#1927995
snsabot: Logged on 2019-08-12 13:29:09 mp_en_viaje: and yes alflog still teletype, http://logs.nosuchlabs.com/log/trilema/2018-04-18#1802801 loaded in ~5s
asciilifeform: http://logs.nosuchlabs.com/log/trilema/2019-08-13#1928201 << i dun 'upgrade', lol. and also admittedly never before tried to get a www to load in <1s specifically. ( btw -- does trilema for all pgs ? )
snsabot: Logged on 2019-08-13 06:02:15 mp_en_viaje: http://logs.nosuchlabs.com/log/trilema/2019-08-12#1928191 << you know, someone fucked something up somewhere, because recall http://btcbase.org/log/2019-07-14#1922400 ? sounds exactly the same, server just sits there for MINUTE before serving anything, for no reason. this didn't use to be the case, i suspect some "upgrade" broke the stack you two both happen to use ?
asciilifeform: if somebody's sat for ~minute~ , sounds like sslism ( not in use on dulap and never will be )
asciilifeform brb, meat
feedbot: http://trilema.com/2019/holy-shit-i-woke-up-by-myself/ << Trilema -- Holy shit, I woke up by myself.
mp_en_viaje: asciilifeform, looks like about 1-2s for a large (400kb) page (a ; b), though id you actually look into the waterfalls the page itself is composed in 400-500ms, the rest being wats for random repeatable elements (which most browsers don't re-load) like avatar.pngs favicon.ico sbtop.png etc
asciilifeform: mp_en_viaje: log now also 1-2s here ( ~3 for the 'heavy' historic pg linked earlier )
asciilifeform: but would like to confirm that other folx see similar figure, and aint an artifact of asciilifeform's local cachism
asciilifeform: mp_en_viaje: ty for the tool link, will try with it
asciilifeform not habitual wwwist and was unaware this existed
mp_en_viaje: $ time curl http://logs.nosuchlabs.com/log/trilema/2019-08-09 yields real 0m5.405s
mp_en_viaje: tbh in your case looks mostly like slow / poorly burstable pipe ; if you look at the waterfal it's mostly grey (receiving), 95% of the 2.x seconds they wait
mp_en_viaje: use the mouse hoover tooltips, they're informative. anyways, the webpage itself is good grade, by web standards.
mp_en_viaje: 3s for 80kb is scandalous, but i'd say it's the poor q of your bw, nothing you can do about that from software.
asciilifeform: mp_en_viaje: hm why graphic browser gives < 1s for same here ?
diana_coman: http://logs.nosuchlabs.com/log/trilema/2019-08-13#1928221 -> I got real0m9.634s
snsabot: Logged on 2019-08-13 10:37:40 mp_en_viaje: $ time curl http://logs.nosuchlabs.com/log/trilema/2019-08-09 yields real 0m5.405s
asciilifeform: perhaps curl dun eat gzip
mp_en_viaje: asciilifeform, possibly because caches.
mp_en_viaje: hit shift-ctrl-r or w/e invalidates all cache in your browser
asciilifeform: knew this, same sans cache
asciilifeform: something is odd, i sat all evening last night reading vintage #t logz via my displayer and none ate >3s
mp_en_viaje: who even knows, maybe "smart router" decided to priviledge you connection
asciilifeform: ( after upped buffer on box and reenabled gzipism )
asciilifeform: mp_en_viaje: possibly i did not elaborate , but on box's own shell, if viewing through ngx , fully 1-2s moar delay than if direct to port where py proggy
asciilifeform: and in fact accts for all but ~70msec of a heavy pg
asciilifeform: i found no obvious cause, but in process of knob turning combinatorics cut the delay somewhat ( seems to only be detectable in graphic wwwtrons tho, and , annoyingly, cannot actually get useful number from those, unlike curl )
asciilifeform: tried large buffers, no buffers, various tcpism knobs, no obv culprit for the impedance mismatch (what i suspect this is )
asciilifeform: pretty certain that the pipe itself is blameless, i get consistent realtime work through it, not once long pings much less subjective feel of 'mars probe'
mp_en_viaje: asciilifeform, do inspect element or similar, developer toolbox, we
mp_en_viaje: all gfx browsers have a timer thing
asciilifeform: was looking at https://gtmetrix.com/reports/logs.nosuchlabs.com/m3to2XTx , somewhat cryptic, will check out others
asciilifeform: iirc chromisms have timer, i have box w/ one
mp_en_viaje: firefox too
mp_en_viaje: unless they took it out, like the screenshot thing
asciilifeform: https://tools.pingdom.com/#5b22d518dfc00000 << this one confirms my init observ., almost whole second in 'connect' + 'wait' ( and 5 in receive , doesn't say in what pattern, but visually curl in the 'slow even on local shell' variant appears 'bursty' )
asciilifeform: http://logs.nosuchlabs.com/log/trilema/2019-08-13#1928224 << ftr i routinely get 1-2 MB/s from my chair to dulap and reverse
snsabot: Logged on 2019-08-13 10:41:12 mp_en_viaje: 3s for 80kb is scandalous, but i'd say it's the poor q of your bw, nothing you can do about that from software.
asciilifeform: it aint best possible pipe, but pretty good; and , as i pointed out already, effect persists ~without~ it ( i.e. operating curl within dulap shaves only 200msec vs to my chair; and has same peculiar 'burst' pattern and wild variance in total time )
asciilifeform: via chromistic timer , get 1.40s for the heaviweight sample
snsabot: Logged on 2018-04-18 21:58:18 mircea_popescu: in unrelated news : i can't recommend the dsc-rx100 quite warmly enough. it's a splendid sony compact with zeiss lenses that's easily the best camera i ever saw.
asciilifeform: this thing dun seem to distinguish the phases of the load tho (only distinguished load from rendering, and cuts the latter into phases. which makes a kind of sense, tool is evidently for debug of 'slow cuz hands of who wrote html-shitter grow from arse' scenarios )
asciilifeform will try on various FF later today. other folx welcome to report theirs.
asciilifeform: if this box weren't also running phuctor, could simply hang the logger on port 80 and get, theoretically, 200-300msec loads, loox like.
asciilifeform: aint nginx, apache, etc supposed to ~cure~ impedance mismatches, rather than cause'em...
asciilifeform: http://logs.nosuchlabs.com/log/trilema/2019-08-13#1928196 << lobbes does this result in same eggogs when eating your archive as before ?
snsabot: Logged on 2019-08-13 01:25:21 feedbot: http://blog.lobbesblog.com/2019/08/znc2tmsr-vpatch/ << lobbesblog -- znc2tmsr vpatch
asciilifeform: btw time curl -H "Accept-Encoding: gzip" http://logs.nosuchlabs.com/log yields avg. 0.4s . so indeed default curl dun eat gzip.
asciilifeform: this aint a cure tho (afaik not all browsers eat it)
asciilifeform: the 'heavy' page -- 1.4-2s avg. on same
asciilifeform: ( from asciilifeform's chair ; 0.9 when executed locally on dulap , via ng i.e. the public url )
asciilifeform: the 'heavy' pg again, via yet another 'www debugger'. 0.5s time-to-first-byte ! ..and defo shows the 'burst' pattern.
asciilifeform: or hm, possibly misread, and it's 0.737s
asciilifeform has nfi why the # given for 'all' vs the 1st 'resource' could differ
asciilifeform: it is even possible that the pipe is to blame, i.e. the measurement is 'random' depending on how congested the pipe at piz on acct of e.g. the multiple trb noads sitting there
asciilifeform: however :
asciilifeform: e.g. time curl "http://ossasepia.com/" > /dev/null >> real 0m1.597s and weighs 108k atm
asciilifeform: so i still suspect it aint the pipe ( diana_coman lives at piz )
asciilifeform: all the clues point to nginx.
asciilifeform: for comparison : diana_coman's www via this same tool .
asciilifeform: for non-piz comparison : trilema www, via same .
asciilifeform now thoroughly stumped, ~all 3~ seem to show ~same 'time to first byte'
asciilifeform: and now phf's log -- of 'heavy' pg -- via same. pretty interesting. total load time (to the tester author's box) -- 0.646s ; 1st byte -- 0.254s .
asciilifeform: i'm beginning to suspect that nginx, apache, etc. are intrinsically dead weight, when sat in front of a programmatic html shitter
asciilifeform: and that phf did not use these, but hung up his logger straight on port80
asciilifeform: attn panzers : whichever 1 of you is searching for 'cuts of the meat' -- that's in pre-dragon log, not yet imported.
feedbot: http://trilema.com/2019/burn-after-reading/ << Trilema -- Burn After Reading
asciilifeform: meanwhile, added 'download daily db snapshot' link to logotron.
asciilifeform will put in readme , for next patch, how to set up the cron for these
asciilifeform: currently set to snapshot at 0 hrs. daily. (snapshot itself takes ~8sec at the current mass of the db)
asciilifeform: presently, weighs (compressed) ~23MB.
asciilifeform: i expect will be roughly 3x that when all the back-archives added.
asciilifeform: if folx want to mirror, i'd prefer that also happen daily, otherwise no amt of pipe will suffice
asciilifeform: this ^ is simple means of testing the logger for folx not yet ready to stand up bot. e.g. : zcat dump.gz | psql -U nsabot -d nsalog will eat (assuming you already inited db w/ the provided schema)
BingoBoingo: Las empresas de consumo masivo aumentan sus precios hasta 15% y el Gobierno evalúa qué hacer con "Precios Esenciales" << Headline across the river
BingoBoingo: Also: "Honda dejará de fabricar autos en la Argentina y se focalizará en la producción de motos"
feedbot: http://qntra.net/2019/08/4-tor-users-sentenced-in-us/ << Qntra -- 4 Tor Users Sentenced In US
asciilifeform: hey BingoBoingo , couldja run for me these timings plz
BingoBoingo: asciilifeform: http://p.bvulpes.com/pastes/kJJ54/?raw=true
asciilifeform: ty BingoBoingo
BingoBoingo: In other news, lower office holding Cristinists are already proposing a return of the Blue Dollar. Nevermind the Argentine Central bank is dumping dollars onto the market now just trying to slow the Peso Argentino's fall.
feedbot: http://pizarroisp.net/2019/08/13/pizarro-isp-july-2019-report/ << PizarroISP -- Pizarro ISP July 2019 Report
asciilifeform: BingoBoingo: let's get the 2nd xyz plant specced an' bought this mo
asciilifeform: if at all possible
asciilifeform: BingoBoingo: considering we're 100% outta unoccupied holes in which to put folx
asciilifeform: the tyre fire of 'hey we've no logger' is , if not put out, then at low smoulder, so let's get this item moving
asciilifeform still ill at ease re logger, pending confirmation that the kit actually stands up somewhere other than in asciilifeform's hands
asciilifeform: see also
snsabot: Logged on 2019-08-11 17:08:50 asciilifeform: diana_coman: my pov, is that would like to get all 'magick' the fuck out of logotronics.
asciilifeform: http://logs.nosuchlabs.com/log/trilema/2019-08-13#1928291 << i take it 'blue dollar' is/was same thing as sovok's 'convertible ruble' ?
snsabot: Logged on 2019-08-13 18:24:25 BingoBoingo: In other news, lower office holding Cristinists are already proposing a return of the Blue Dollar. Nevermind the Argentine Central bank is dumping dollars onto the market now just trying to slow the Peso Argentino's fall.
asciilifeform: ( y'know, the one that was Officially Pegged (tm) at 15 to 1 $ , whereas in reality 100+:1 )
BingoBoingo: asciilifeform: iirc "Blue" was the dollar that tracked actual supply/demand
BingoBoingo: <asciilifeform> BingoBoingo: let's get the 2nd xyz plant specced an' bought this mo << Aite
asciilifeform: hm i thought that was the only remaining kind after dethronement of whore