00:02:14jacksonchen666 (jacksonchen666) joins
00:09:00<icedice>Ryz: Until the hosting costs or the CSAM reports makes the image host shut down, more or less
00:09:23<icedice>One of the two is usually what shuts down an image host
00:24:22etnguyen03 (etnguyen03) joins
00:25:55Dallas quits [Ping timeout: 272 seconds]
00:34:30BearFortress quits [Client Quit]
00:35:41BearFortress joins
00:40:02FireFly quits [Remote host closed the connection]
00:42:34FireFly joins
00:43:02<Ryz>Mmm, interesting insight fro you icedice o#o;
00:43:19<Ryz>Another one dead: https://web.archive.org/web/20231012003512/https://pngkit.com/
00:43:44<Ryz>Unsure to what extent on grabbing these if perception-wise, there's a low amount of original content~ o.o;
00:44:13<Ryz>This one isn't dead though: https://www.pngkey.com/
00:44:38<icedice>tbh, that data is more from file hosts than it is from image hosts now that I think about it
00:44:39<joepie91|m>I think seekpng was a scraper?
00:45:16<icedice>Though a lot of the ones that die are Pomf-based
00:45:19<icedice>So they're both
00:45:31<joepie91|m>a cached version of seekpng has 'nicepng' in the copyright footer
00:45:35<joepie91|m>and nicepng seems still up
00:45:44<joepie91|m>with noticeably similar content
00:46:49<icedice>AnonFiles was killed by the latter, for example
00:47:26<icedice>Svea Hosting had enough with all the shit people were posting on there via Tor so they pulled the plug
00:48:49<icedice>AnonFiles had set up proxy.is in the Seychelles as a front company when they hosted their own reverse proxy service
00:49:11<joepie91|m>... do I know Svea Hosting under some other name?
00:49:20<icedice>svea.net
00:49:31<icedice>Obehosting reseller
00:49:33<joepie91|m>besides that I mean
00:49:35<joepie91|m>they look familiar to me but I don't recognize the name
00:49:53<icedice>They host some filehosts and probably some Tor exit nodes
00:50:03<icedice>If it's legal in Sweden, they'll host it
00:50:26<joepie91|m>for a price, it looks
00:50:29<FireFly>I mean the wobsite has generic template vibes tbf
00:50:33<joepie91|m>129 EUR/month for an rpi :)
00:50:40<joepie91|m>FireFly: I mean familiar in terms of pricing and service offerings
00:50:41<icedice>Sweden is expensive as shit
00:50:47<FireFly>ah
00:50:47<icedice>Duopoly there iirc
00:50:49<joepie91|m>certainly not that expensive
00:50:59<icedice>Cheapest you can get in Sweden is Portlane
00:51:00<FireFly>duopoly in what domain?
00:51:07<FireFly>ah transit?
00:51:12<FireFly>hm no idea
00:51:24<joepie91|m>speaking of, what happened to gridlane?
00:51:37<icedice>iirc there are like two companies that operate most of the fiber lines
00:51:57<icedice>Though I might be confusing it with Finland
00:52:07<icedice>It's like that there
00:52:14<joepie91|m>I've definitely seen cheaper providers than this in Sweden
00:53:44<icedice>Portlane is cheap for co-location iirc
00:55:08<FireFly>dunno really how it works with IXes and stuff, but otherwise I think there's a couple companies with fibre networks and a lot of areas also have municipal fibre networks
00:55:19<FireFly>so I think it's not too bad :o at least for consumers
00:59:47<icedice>I wonder if it was submarine cables I was thinking about
01:00:01<icedice>My memory is pretty foggy on it
01:00:59<Ryz>Another one, https://www.pngfind.com/ - probably will stop posting; I just seriously ponder, this particular iteration of how these things crop up
01:04:08<Ryz>Unless somehow people are interested in me keeping finding them somehow, somewhere, if only to find countless wormholes of these things
01:04:19<icedice>Seems like a lot of the submarine cables are owned by Aurelion (Telia) and Elisa, so it was probably that
01:05:04<icedice>There are three other companies that operate submarine cables to/from Finland though and one of them seems to have a few
01:05:12<icedice>So I guess the situation has gotten better
01:05:31<icedice>https://www.submarinecablemap.com/
01:22:47HP_Archivist (HP_Archivist) joins
01:26:45HP_Archivist quits [Client Quit]
01:35:16HP_Archivist (HP_Archivist) joins
01:38:01<nulldata>https://www.bleepingcomputer.com/news/security/cloudflare-dashboard-and-apis-down-after-data-center-power-outage/
01:38:46<nulldata>"We operate in multiple redundant data centers in Oregon that power Cloudflare’s control plane (dashboard, logging, etc). There was a regional power issue that impacted multiple facilities in the region. The facilities failed to generate power overnight. Then, this morning, there were multiple generator failures that took the facilities entirely
01:38:46<nulldata>offline," the spokesperson said.
01:55:26AlsoHP_Archivist joins
01:59:04HP_Archivist quits [Ping timeout: 265 seconds]
02:05:36Hackerpcs quits [Client Quit]
02:07:38Hackerpcs (Hackerpcs) joins
02:11:18<fireonlive>so i got an alert for 'archive team' today... "The First Annual Archive Team Group Getaway To Put Jon In His Place by paperdream"
02:11:21<fireonlive>"Jon is in a compromising position, but luckily his assistants don't know it's him. A few days later, his assistants definitely know it was him. Instead of letting it impact their work relationship, though, they make a proposition. What follows are two weeks of fun, friendship, and triple-teaming their boss until he can't think straight.
02:11:21<fireonlive>Platonically, though. Definitely not romantically, everyone promises."
02:11:34<fireonlive>NSFW; https://archiveofourown.org/works/47907019/chapters/120783043
02:12:09<nicolas17>:|
02:12:37<fireonlive>the tags are something
02:13:41<fireonlive>i didn't know we had so much fun here
02:19:51AlsoHP_Archivist quits [Ping timeout: 265 seconds]
02:23:43etnguyen03 quits [Ping timeout: 272 seconds]
02:31:18<@JAA>I must've missed the invitation. :-(
02:33:10<fireonlive>:(
02:43:07<nulldata>Too Hot For Archive
02:43:38etnguyen03 (etnguyen03) joins
03:00:22pseudorizer quits [Client Quit]
03:00:27etnguyen03 quits [Ping timeout: 265 seconds]
03:01:05pseudorizer (pseudorizer) joins
03:26:45etnguyen03 (etnguyen03) joins
03:43:42<fireonlive>i need like a list of my channels now x_x
03:43:51<fireonlive>send people to a website or something lol
03:44:58<@JAA>Have a single dispatch channel. When someone joins, they get invited into all the other channels and then kicked by eggdrop or something like that.
03:45:04<project10>^
03:45:19<nicolas17>:|
03:45:22<@JAA>Then people see the invites and can choose which of them to follow.
03:45:58<@JAA>Or just put the list in the topic of #fire-spam and point everyone there elsewhere. :-P
04:07:00<fireonlive>oooh mass invite is a neat idea :3
04:27:35icedice quits [Client Quit]
04:59:53<fireonlive>https://www.404media.co/in-defense-of-ram-on-apple-silicon/
05:00:01<fireonlive>for once, i am not on daddy cookie's side on this
05:02:07<@JAA>Whaaat‽
05:02:38<fireonlive>i know, i know... i can barely believe it myself :'(
05:02:39<@JAA>On another note, if anyone needs a time machine, the author of that article appears to have one:
05:02:43<@JAA>> Two years ago, I got a 14” MacBook Pro with an M2 Pro processor
05:03:04<@JAA>M2 Pro was only released in January.
05:03:14<fireonlive>lmao
05:09:29DogsRNice quits [Read error: Connection reset by peer]
05:11:36<pabs>-rss/#hackernews- HTTP 999 – the unofficial status code: https://http.dev/999 https://news.ycombinator.com/item?id=38127131
05:11:42<nicolas17>"Man mistakes RAM for unified memory" what
05:12:04<nicolas17>with unified memory you need *more* RAM since you can't say "yeah I have 8GB main memory but my GPU also has 4GB of its own"
05:13:28<@JAA>pabs: Yeah, they've been doing that for years.
05:23:19etnguyen03 quits [Client Quit]
05:50:31HackMii_ quits [Remote host closed the connection]
05:50:31sec^nd quits [Read error: Connection reset by peer]
05:50:54sec^nd (second) joins
05:51:08HackMii_ (hacktheplanet) joins
06:08:05<fireonlive>https://xkcd.com/2850/
06:20:19hitgrr8 joins
06:26:33monoxane5 (monoxane) joins
06:28:49monoxane quits [Ping timeout: 272 seconds]
06:28:49monoxane5 is now known as monoxane
06:30:16<nicolas17>maybe I should have used line-delimited json instead of putting separate json files into a .tar x_x
06:30:41<nicolas17>importing these tars into a database is taking forever and I'm not sure if the bottleneck is zstd, tar, json parsing, or sqlite
06:31:53<nicolas17>probably python tarfile is too dumb and slow...
06:56:53<qwertyasdfuiopghjkl>https://blog.cloudflare.com/post-mortem-on-cloudflare-control-plane-and-analytics-outage/ https://news.ycombinator.com/item?id=38138640
06:59:08<fireonlive>thanks qwerty!
06:59:11<nicolas17>:popcorn:
06:59:32<fireonlive>get your 🍿🍿🍿🍿🍿🍿🍿🍿🍿🍿 here!
07:14:30<nicolas17>...wow this is 402GB of JSON
07:14:40<nicolas17>or rather
07:14:55<nicolas17>402GB of tars containing JSON (tar adds a lot of padding)
07:16:36<nicolas17>zstd compressed it to 22GB
07:16:55project10 pets zstd
09:02:28qwertyasdfuiopghjkl quits [Ping timeout: 261 seconds]
09:07:50qwertyasdfuiopghjkl (qwertyasdfuiopghjkl) joins
09:46:37benjinsm joins
09:46:49benjins quits [Remote host closed the connection]
09:46:51Arcorann quits [Remote host closed the connection]
09:46:51nic9 quits [Client Quit]
09:47:13nic9 (nic) joins
09:52:29Arcorann (Arcorann) joins
10:16:56jacksonchen666 is now known as RJHacker27920
10:17:00jacksonchen666 (jacksonchen666) joins
10:19:11RJHacker27920 quits [Ping timeout: 245 seconds]
11:34:07pseudorizer quits [Client Quit]
11:35:05pseudorizer (pseudorizer) joins
11:54:14jacksonchen666 quits [Remote host closed the connection]
11:54:58jacksonchen666 (jacksonchen666) joins
12:15:59benjinsmi joins
12:18:04HP_Archivist (HP_Archivist) joins
12:19:41benjinsm quits [Ping timeout: 272 seconds]
12:30:14etnguyen03 (etnguyen03) joins
13:00:59benjinsm joins
13:01:03AlsoHP_Archivist joins
13:01:08gfhh1 joins
13:04:37HP_Archivist quits [Ping timeout: 265 seconds]
13:04:37gfhh2 quits [Ping timeout: 265 seconds]
13:05:35benjinsmi quits [Ping timeout: 265 seconds]
13:05:35Arcorann quits [Ping timeout: 265 seconds]
13:06:37icedice (icedice) joins
13:06:48Arcorann (Arcorann) joins
13:14:46Arcorann quits [Ping timeout: 265 seconds]
13:19:18AlsoHP_Archivist quits [Client Quit]
13:19:39HP_Archivist (HP_Archivist) joins
13:35:57jacksonchen666 is now known as RJHacker57264
13:36:01jacksonchen666 (jacksonchen666) joins
13:36:16RJHacker57264 quits [Ping timeout: 245 seconds]
13:39:52benjins joins
13:41:30HP_Archivist quits [Remote host closed the connection]
13:41:30gfhh1 quits [Remote host closed the connection]
13:41:30icedice quits [Remote host closed the connection]
13:41:49gfhh1 joins
13:41:49HP_Archivist (HP_Archivist) joins
13:41:50icedice (icedice) joins
13:43:46benjinsm quits [Ping timeout: 260 seconds]
13:46:51gfhh2 joins
13:47:06Miori8 joins
13:47:08girst_ (girst) joins
13:47:14Sanqui_ joins
13:47:16Sanqui_ quits [Changing host]
13:47:16Sanqui_ (Sanqui) joins
13:47:16@ChanServ sets mode: +o Sanqui_
13:47:18monoxane2 (monoxane) joins
13:47:19franga20004 joins
13:47:40balrog_ (balrog) joins
13:47:43Ryz quits [Client Quit]
13:47:43@Sanqui quits [Client Quit]
13:47:43Miori quits [Client Quit]
13:47:43TheTechRobo quits [Client Quit]
13:47:43imer quits [Client Quit]
13:47:43that_lurker quits [Client Quit]
13:47:43franga2000 quits [Client Quit]
13:47:43s-crypt2 quits [Client Quit]
13:47:43emberquill08 quits [Client Quit]
13:47:43nic9 quits [Client Quit]
13:47:43balrog quits [Client Quit]
13:47:43monoxane quits [Client Quit]
13:47:43HP_Archivist quits [Remote host closed the connection]
13:47:43icedice quits [Remote host closed the connection]
13:47:43girst quits [Client Quit]
13:47:43gfhh1 quits [Remote host closed the connection]
13:47:43wickedplayer494 quits [Remote host closed the connection]
13:47:43Miori8 is now known as Miori
13:47:43franga20004 is now known as franga2000
13:47:43girst_ is now known as girst
13:47:43monoxane2 is now known as monoxane
13:47:46icedice (icedice) joins
13:47:48emberquill08 (emberquill) joins
13:47:49nic9 (nic) joins
13:47:49HP_Archivist (HP_Archivist) joins
13:47:51s-crypt2 (s-crypt) joins
13:47:51that_lurker (that_lurker) joins
13:47:55imer (imer) joins
13:48:04wickedplayer494 joins
13:48:14TheTechRobo (TheTechRobo) joins
13:48:15Ryz (Ryz) joins
14:05:31etnguyen03 quits [Ping timeout: 265 seconds]
14:14:36etnguyen03 (etnguyen03) joins
14:17:27gfhh joins
14:20:01gfhh2 quits [Ping timeout: 272 seconds]
14:31:25etnguyen03 quits [Ping timeout: 272 seconds]
14:33:41decky joins
14:37:07decky_e_ quits [Ping timeout: 272 seconds]
14:42:01jacksonchen666 quits [Remote host closed the connection]
14:42:28jacksonchen666 (jacksonchen666) joins
14:50:26jacksonchen666 quits [Remote host closed the connection]
14:50:54jacksonchen666 (jacksonchen666) joins
15:11:25jacksonchen666 quits [Remote host closed the connection]
15:12:32jacksonchen666 (jacksonchen666) joins
15:15:07katocala quits [Ping timeout: 272 seconds]
15:16:02katocala joins
15:23:25HP_Archivist quits [Client Quit]
16:01:05etnguyen03 (etnguyen03) joins
16:29:51_Dango360 quits [Ping timeout: 272 seconds]
16:31:29katocala quits [Ping timeout: 265 seconds]
16:31:47katocala joins
16:32:55HP_Archivist (HP_Archivist) joins
16:35:40Craigle quits [Client Quit]
16:37:55Craigle (Craigle) joins
16:50:45etnguyen03 quits [Ping timeout: 272 seconds]
16:56:26DogsRNice joins
17:00:53Earendil7 quits [Ping timeout: 272 seconds]
17:06:56igloo22225 quits [Remote host closed the connection]
17:09:13Craigle quits [Client Quit]
17:09:53Craigle (Craigle) joins
17:11:05Earendil7 (Earendil7) joins
17:15:15igloo22225 (igloo22225) joins
17:23:41katocala quits [Ping timeout: 272 seconds]
17:23:57katocala joins
17:44:57katocala quits [Ping timeout: 265 seconds]
17:45:13katocala joins
17:50:04DogsRNice quits [Remote host closed the connection]
17:50:04icedice quits [Remote host closed the connection]
17:50:04gfhh quits [Remote host closed the connection]
17:50:04decky quits [Remote host closed the connection]
17:50:06DogsRNice joins
17:50:22gfhh joins
17:50:22decky joins
17:50:38icedice (icedice) joins
17:51:00DogsRNice_ joins
17:55:35DogsRNice quits [Ping timeout: 265 seconds]
18:13:21<@JAA>That post-mortem just keeps getting worse. lol
18:33:16_Dango360 joins
18:36:54etnguyen03 (etnguyen03) joins
18:47:36<kiska>tldr CF post mortem?
18:53:41<@JAA>A Flexential DC in Oregon lost one of its power grid connections, then a ground fault brought down the other one plus the diesel generators. The UPS ran empty much quicker than anticipated, and the emergency systems for bringing the generators back up weren't connected to it. Some of Buttflare's systems weren't redundant with the other two DCs in the area they use, and some of their redundant systems
18:53:47<@JAA>depend on non-redundant ones. They switched over to DR infra in Europe, which was crushed by the thundering herd and didn't support all services properly.
19:18:17lukash99 joins
19:18:57lukash9 quits [Ping timeout: 272 seconds]
19:18:57lukash99 is now known as lukash9
20:13:38<nicolas17>ok looks like I have 258GB of JSON and the rest is tar overhead lol
20:20:11magmaus3 quits [Client Quit]
20:20:34magmaus3 (magmaus3) joins
20:21:13jacksonchen666 quits [Remote host closed the connection]
20:21:14HackMii_ quits [Write error: Broken pipe]
20:21:40jacksonchen666 (jacksonchen666) joins
20:21:45HackMii_ (hacktheplanet) joins
20:26:54<nicolas17>and my DB imported failed after like 3 hours due to a dumb bug
20:28:30<nicolas17>>if the JSON file doesn't contain "software", count it as "skipped early" (for statistics) and skip it
20:28:31<nicolas17>>do the actual JSON parsing
20:28:33<nicolas17>>if the parsed dict["kind"] isn't equal to "software", count it as "skipped late" [I forgot the actual "and skip it" here x_x]
20:28:34<nicolas17>>load it into the database
20:28:36<nicolas17>it reached https://music.apple.com/us/album/software/1552564176?i=1552564177&uo=4 and crashed because it didn't have the properties expected for an app (because it's not an app)
20:28:43<joepie91|m>JAA: that 'redundancy' is absolute clowncar shit, lol
20:28:53<joepie91|m>"oh we thought we had HA but we never actually tested it"
20:29:10<@JAA>Yeah
20:29:33<@JAA>And then 'oh, this is HA, and we tested it, but it depends on non-HA stuff we didn't test, oops'.
20:30:13<joepie91|m>no end-to-end service dependency graph either
20:30:19<joepie91|m>no 24/7 experienced tech availability
20:30:50<nicolas17>"no 24/7 experienced tech availability" that part is on the DC
20:30:51<@JAA>Same goes for Flexential, apparently. I wonder if they'll publish a post-mortem, too.
20:31:17<joepie91|m>what is this, fucking Joe's Datacenter?
20:31:37<katia>burst.net
20:31:42<joepie91|m>nicolas17: well no because that's something you negotiate in your contract and *pay for* when you are running critical services like cloudflare does with their control plane
20:31:54<joepie91|m>(and verify that it actually happens)
20:32:13<joepie91|m>not having 24/7 coverage on a network edge is one thing, but come on, this was their control plane
20:33:29<joepie91|m>there's more stuff in their postmortem that raises serious questions about what cloudflare's contract looked like
20:33:30<nicolas17>I'm sure stuff like "These batteries are sufficient to power the facility for approximately 10 minutes. That time is enough to bridge the gap between the power going out and the generators automatically starting up." is in the contract and paid for too, and yet... :P
20:34:09<@JAA>4 and 10 are the same order of magnitude, so it's fine, right? :-P
20:34:11<joepie91|m>like how they complain about not being notified of the power change, but they talk about 'best practice' rather than 'per contractual agreement'. does that mean that it wasn't in the contract?
20:34:33<nicolas17>and "Flexential's access control system was not powered by the battery backups, so it was offline" sounds like stuff that shouldn't need to be clarified in a contract to be done right
20:34:50<nicolas17>but yes I'm sure there's a ton of things where Cloudflare *should* have had it in the contract and didn't
20:35:01<joepie91|m>oh I'm sure that Flexential made errors as well
20:35:24<joepie91|m>but there are also a number of things that CF is pawning off on Flexential as if they hold no responsibility for it themselves, when they absolutely should have known better
20:35:40<@JAA>If Buttflare had properly set up their system to tolerate a full DC outage as they originally intended, everything would've been fine.
20:35:55<@JAA>If Flexential had properly operated their DC, everything (probably) would've been fine.
20:36:00<@JAA>So yeah, both suck.
20:36:04<nicolas17>everyone seems to be hit by unexpected dependencies
20:36:04<joepie91|m>like, for this sort of service I expect every I to be dotted and every T to be crossed, in minute detail, with several reviews, given the sheer importance of the system
20:36:10<joepie91|m>and that threshold was very clearly not met
20:36:29<fireonlive>looool burst.net
20:36:30<nicolas17>yesterday's github outage:
20:36:36<nicolas17>"Rolling back the authorization service change was delayed as parts of the deployment infrastructure relied on the authorization service and required manual intervention to complete"
20:37:09<nicolas17>Cloudflare's other outage a few days ago: "When this happened, Workers KV became unreachable in production. Although automated alerting detected the issue immediately, there was a delay [until] were actually able to perform the roll back. This was caused by the fact that multiple tools within Cloudflare rely on Workers KV."
20:37:48<joepie91|m>"we don't need network diagrams!"
20:37:57<nicolas17>an AWS outage a few years ago: "This congestion immediately impacted the availability of real-time monitoring data for our internal operations teams, which impaired their ability to find the source of congestion and resolve it."
20:37:59<joepie91|m>"we're an agile tech company!"
20:38:10<joepie91|m>"oops how come we didn't know about this dependency"
20:45:51<@JAA>> YouTube's plan backfires, people are installing better ad blockers
20:45:59<@JAA><surprised_pikachu.png>
20:46:31<nicolas17>link?
20:46:35<@JAA>https://www.androidauthority.com/youtube-ad-block-installs-3382289/
20:47:56<katia>i've never seen the banner
20:48:08<katia>firefox developer edition + ublock origin
20:48:51<@JAA>Me neither, but I only watch videos with mpv + yt-dlp anyway, never in the shitty web interface.
20:57:08BlueMaxima joins
21:57:54lunik173 quits [Client Quit]
21:58:23lunik173 joins
22:04:54<nicolas17>holy crap, I tried parsing every json file in the tar and dumping it again as a single line + '\n' (ie. turning tarball of json files into a single ndjson), and the ETA is 40 hours
22:05:00<nicolas17>is Python's json module written in pure Python?
22:06:52<@JAA>Yes
22:07:29<@JAA>There are drop-in alternatives written in C. I happen to maintain one, ultrajson.
22:09:27<nicolas17>I happen to already have ujson installed as a dependency of something else
22:09:55<nicolas17>ETA 6 hours...
22:10:18<@JAA>:-)
22:19:40<nicolas17>hmmm I think the official json module actually does 8 hours, my ETA was just crap because I was counting tarballs instead of json files, and the number of jsons in each tar varies a lot
22:20:56<nicolas17>or 10 hours... but still not the 40 I got before
22:41:32benjinsm joins
22:43:42benjinsmi joins
22:46:03benjins quits [Ping timeout: 272 seconds]
22:46:33benjinsm quits [Ping timeout: 265 seconds]
22:56:20<nicolas17>JAA: I missed the obvious solution in this particular case -.-
22:56:25<nicolas17>don't use a JSON parser
22:56:29<nicolas17>just... delete the newlines
22:57:07<@JAA>Well, assuming your JSON is well-formed.
22:57:19<@JAA>I've seen JSON with literal LF in strings and stuff like that.
22:57:20<nicolas17>it came out of json.dump()
22:57:28<nicolas17>should be fine
22:57:37<@JAA>Ah yeah, then it should probably be fine.
22:58:12_Dango360 quits [Read error: Connection reset by peer]
23:02:02Dango360 (Dango360) joins
23:21:28hitgrr8 quits [Client Quit]
23:27:07<fireonlive>"Lawina Extreme Movie Blog Blog Archive Team Russia Cloudy Girl Pics - xxxbullet.com" we are really expanding out lately
23:36:19DigitalDragons quits [Client Quit]
23:36:43DigitalDragons joins