Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
Posted on 20-08-25, 23:22 in game genie "special load" option...
Dinosaur

Post: #761 of 1285
Since: 10-30-18

Last post: 16 days
Last view: 2 hours
Wonder if Higan's manifest system can help emulating such ridiculousness (disconnecting vital pins at will)

But there are easier ways to achieve blank screens, like:
- Loading corrupted/invalid ROMs
- Not loading a game at all
- Load a piece of homebrew whose only purpose is to crash the system right at boot

Licensed Pirate® since 2006, 100% Buttcoin™-free, enemy of All Things JavaScript™
Posted on 20-08-26, 20:15 in game genie "special load" option...
Dinosaur

Post: #762 of 1285
Since: 10-30-18

Last post: 16 days
Last view: 2 hours
I'm tempted to test it, but my Genesis is currently stashed away on a box, which has a pile of boxes over it.

But then, Sega actually licensed the goddamned thing, we didn't had pesky lockout chips, etc. The SNES one most likely will not even boot as it piggybacks on the lockout chip of the plugged-in game.

Licensed Pirate® since 2006, 100% Buttcoin™-free, enemy of All Things JavaScript™
Posted on 20-08-28, 00:21 in Mozilla, *sigh* (revision 4)
Dinosaur

Post: #763 of 1285
Since: 10-30-18

Last post: 16 days
Last view: 2 hours
Regarding the ongoing Fireocalypse in Android situation (Fennec v68 -> Fenix v79):

As you may already be aware, the Chromification of Mozilla is going full throttle (since they DIDN'T sacked the UXtarded art school dropouts UX visionaries). The new "hotness" has the codename of Fenix (that's "phoenix" in Spanish, how ironical), while the former product you and me were still using on our crippled pocket computers was known as Fennec (a species of small fox, but then a "firefox" is a red panda...).

Fenix was actually being rolled out in stages, culminating with the final master update a few days ago for all users on supported Android versions (which for Fenix means Lollipop/5.0+ and latter - KitKat is no longer supported, and if your device can only run that, chances are Fennec already ran like ass on it), and given the current trends on user interface design across the software industry, people hate the forced dumbing-down and are already complaining loudly all around the Internets. Here's Mozilla update FAQ for Fenix, complete with the expected "We're sorry to hear you hate change, so we will pretend we're listening while we actively ignore you!". Oh, and this pearl that really makes my blood boil:
Primary Password is replacing Master Password. Firefox is removing terminology from the browser that has been identified as derogatory or exclusionary. Learn more about why we’re making this change.

...AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAGGGGGGGGGGGGGGGHHHHHHHHHHHHH!!!!!
Fucking SJWs, stop trying to rewrite the language!!!!
Yup, the SJW crybabies brigade also survived the latest Mozilla purge as they're a core part of the current Chromification-as-a-service bullshittery.

Anyway, if this situation looks too familiar to you, it's because we've already been there plenty of times. Sadly there is no Seamonkey/Waterfox/Pale Moon we can defect to, so our options... well, suck:

- Stick to 68.11.0 forever, which is already deemed "DANGER MINES!!!!"
- F-Droid has a "sanitized, 100% free" build of Fennec that closely matched upstream releases (a-la Iceweasel). Except that that one is already doomed as Fennec is no more, and even if you wanted a lobotomy to follow the progress of Fenix in a more "free" way, the build system of Fenix relies on quite a few proprietary bits plus a Mozilla-specific Maven repo which F-Droid refuses to trust, which obviously are a no-go for any potential F-Droid release. Unfortunately for F-Droid maintainers, they really have no choice: it's suck it up and patch up the suckage to de-blobify, or leave the repository without an up-to-date web browser (note that they don't offer Chromium-based web browsers - it's Firefux or nothing!). BTW: shame on you, F-Droid, for using the forum software ABOMINATION known as Dischorse Dicksores Discourt Discourse! My Seamonkey cries HARD every time I accidentally land on a Dicksauce-powered board!
- Consider a non-Gecko way out... which sadly means cementing the Blink monoculture in place. ALL HAIL OUR NEW BROWSER FUHRER, THE ETERNAL CHROME 4LYFE!
- Some poor, desperate, solitary soul is joining on the insanity that implies pulling a Pale Moon over Fenix (the new core with a hacked-up UI to undo some of the UXtards sabotage). Good luck, as this guy (or gal) is seriously going to need all he/she can get!

Needless to say, v68.11.0 is my final web browser for cellphones as I'm not upgrading, ever. Thankfully I don't use Google Play, Aurora can't really do autoupdates on its own (which is GOOD), and I don't Internet from my cellphone, except for utterly menial tasks like knowing the current exchange rates or some Wikipedia whenever I'm on the go. For everything else, let this be your umpteenth reminder that cellphones are NOT replacements for Real Personal Computers™!

---

Speaking about Pale Moon, I tried to give it a chance. AGAIN. Honestly, for real this time. This Moonchild guy really thinks a lot like me: we both hate the current incarnation of Mozilla, the horrible "cellphones-first-and-only" design culture that it's ruining the Internet and computers in general, software bloat in general, and the whole "all your base belongs to Microsoft Google, OR ELSE!" situation. Unfortunately no two great minds really think alike, and after taking a brief look at Pale Moon forums, I remember why I'm not letting Pale Moon enter my computers, ever. No, it's not due to broken web sites whose shitty developers only care about Google Chrome failing to work at all on Pale Moon. Nope, it goes deeper than that:

- The whole "PDFs don't belong to web browsers, You're Doing It Wrong!" attitude against pdf.js (I've heard someone came up with an PM-specific pdf.js addon)
- The *BSD port fiasco ("You must...", which also reminds me that PM fanboys "contributors" can be as outrageously annoying as Moonchild himself, if not more)
- They claim they listen to their users, getting rid of features noone wants. This is fair, except when said features are functionality people require to forcefully interact with webshit on which they have no options (aside of cellphones, or if you're lucky, a Metro UWP crapp), like WebRTC. This is not DRM (the world will be a better place without Hollywood), this is a feature people need for videoconferencing without having to install a Electron crapp that might not even exist for their platform of choice! (*cough*Linux*cough*) Normally, in the pre-Covidiots age I would have said "fuck WebRTC", but we're in the new normal, and people need to get shit done even if the only tool they have at their reach is... well, a web browser (remember: we're supposed to believe that the browser is the new OS, just like Linux is only a kernel). Most (if not all) mentions of said technological term end in confrontations or derisive answers.
- THIS. Wow, so we're talking about "doing illegal things" on a thread where people is trying to block ads on an ad-funded commercial video site (an action that it's clearly frowned upon on the ToS you don't read, but nope, every user of youtube-dl is a frickin' criminal whose only goal in life is to pirate everything, because noone ever thought of the idea that videos actually work better on real media player applications, and you can indeed pipe streaming video from The Tubes to your favorite player without a single bit of it touching your HDD platters/SSD cells, therefore respecting the spirit of said ToS of "thou shall not copy that floppy!". Give me a fucking break, you double-standard assholes!
- Actively discouraging people from going deeper, messing with the code, or in general getting involved if their ultimate goals don't exactly align with their one-and-only vision. Isn't it free and open source software then?!
- The highly novice-unfriendly attitude of Moonchild and friends towards new users that prefer to open a new thread (your hosting isn't charging you by the thread, right?) instead of getting lost in a forum search that half of the times fail to yield useful results.
- The anti-Noscript saga ("You're not smart enough to deal with it, therefore it's BANNED FOR EVERYBODY!")
- I could go on and on and on...

So yeah, it's Seamonkey for me 'till the very end. And on cellphones, I'll be backing up that precious Fennec .APK for posterity.


BONUS LINK: When Poetteringware hateboners clash with The Moonchild Way®.
Turns out that this Matt Tobin guy is an even bigger asshole than Moonchild himself: "do what I say OR ELSE!".
Another reason to keep Pale Moon the hell away from your *nix distro repositories, because they don't really want users anyway. I can understand the branding restrictions (after all, Mozilla has those too), but the way THIS GUYtries to enforce the rules is abhorrent, to say it politely.

Licensed Pirate® since 2006, 100% Buttcoin™-free, enemy of All Things JavaScript™
Posted on 20-08-28, 14:18 in Upcoming game announcements/news (revision 1)
Dinosaur

Post: #764 of 1285
Since: 10-30-18

Last post: 16 days
Last view: 2 hours
Sega finally figured out how to sell puyos to westerners: by addinng tetrominoes.

There were doubts with the formula on the original Puyo Puyo Tetris when it was launched overseas ~4 years ago, but aside of the kinda limited platform selection (no 3DS, no Xbone, PC got the Denuvo disease*), it seems the bet paid off handsomely. Not only we got Puyo Puyo eSports Champions over here (even on PC!), Sega has just announced there will be a Puyo Puyo Tetris 2 later this year.

Yes, looks like it will be a global launch, in what it seems to be a first for the puyos.
It will be available for all current and next-gen consoles, this time including Xboxes.
And yes, THERE WILL BE A PC VERSION ON STEAM!

*Come on Sega, you've got rid of the Denuvo pest on pretty much all of your games EXCEPT for PPT!? Even Sonic Mania got disinfected eventually... Hope you don't make the same silly mistake with PPT2, because the last thing a fine game deserves are review bombings.

Licensed Pirate® since 2006, 100% Buttcoin™-free, enemy of All Things JavaScript™
Posted on 20-09-05, 14:02 in Computer Hardware News
Dinosaur

Post: #765 of 1285
Since: 10-30-18

Last post: 16 days
Last view: 2 hours
I'm not sure about you guys, but I'm not really trusting my long-term, offline backups to what is essentially tiny, tiny capacitors just yet.

Backups ARE still a thing, yo. On consoles and "cloud-connected" toys devices, backups are Not Our Problem™, but on Real Computers where we do actually create and hoard important stuff, magnetic storage will still be with us for at least a few decades to come. Not all workloads are "gotta go FAAAAAAAAAAASSSSSSSSSST!!!" - a SSD is a must for your OS/applications nowadays, but for videos, photos, music, and archived media, I think we can survive the once-in-a-while performance hit of spinning rust.

If anything, I would want a revival of optical media, if only because of simplicity. Too bad Blu-Ray died on the PC space mostly because of DRM and greed. Where are those 1TB "holographic" discs we were supposed to be using to store our backups nowadays!?

Licensed Pirate® since 2006, 100% Buttcoin™-free, enemy of All Things JavaScript™
Posted on 20-09-06, 14:42 in Computer Hardware News (revision 1)
Dinosaur

Post: #766 of 1285
Since: 10-30-18

Last post: 16 days
Last view: 2 hours
My prediction: SATA ports will still be going strong in 2030 in over half of the planet populated by not-wealthy people.
I still saw new PCs shipping with floppy drives as late as 2007, and PATA ports on motherboards as late as 2014 (albeit behind PATA-to-SATA bridges - ditto for the very last PATA drives ever made which resembled early SATA drives by the use of a bridge chip too).

Also: what stops spinning rust from switching to NVMe? That was what SATA Express was for. But then, I only saw WD demoing SATA Express/NVMe HDDs, and none of those shipped anyway.

And, why people is so hellbent on killing SATA plugs solely over physical dimensions!? It isn't like motherboards are so space-starved - SATA plugs are so tiny you can shoehorn those nearly everywhere, unlike the monstrosity of PATA connectors. Not everybody is using a RPi or a portless razor-thin Mac, y'know...

Licensed Pirate® since 2006, 100% Buttcoin™-free, enemy of All Things JavaScript™
Posted on 20-09-06, 14:46 in I have yet to have never seen it all.
Dinosaur

Post: #767 of 1285
Since: 10-30-18

Last post: 16 days
Last view: 2 hours
Faster Reading From /dev/zero With Linux 5.10

Because your nulls gotta go fast!

Licensed Pirate® since 2006, 100% Buttcoin™-free, enemy of All Things JavaScript™
Posted on 20-09-08, 14:20 in Computer Hardware News
Dinosaur

Post: #768 of 1285
Since: 10-30-18

Last post: 16 days
Last view: 2 hours
Either spinning rust goes full native USB on both 2.5" and 3.5" (the latter won't be happening at least until USB can supply 12V - there will always be a SATA bridgeboard somewhere), or we leave SATA ports alone.

If performance of HDDs eventually bottleneck SATA/6Gb (which may or may not happen in our lifetimes considering that the tech is currently as cutting-edge as it is physically possible), why not just raise clock speeds again? It worked fine for the SAS guys (they do have SAS/12Gb, although by then 10/15K screamers were already passé, NVMe happened, and in the case of slow, bottomless tanks they were just fine with cheap SATA drives as RAID is still a thing)

Posted by wertigon
Posted by kode54
Fine, you buy your new shit, I'll keep using something that still works for another decade.


You are aware single-core computers will not exist anywhere on the market in 2030 right? Maybe for small, low-end embedded systems, but otherwise most of the world will have moved on. Today high-end CPUs starts at 16c32t beasts all the way to 64c128t, this will probably be at *least* 256c/1024t machines by 2030 running at 4-5 GHz boost clocks. Lowest end will probably be 16 cores by then.

At that point, lending one core to focus 100% on a single data transfer OP will be perfectly acceptable even for a Low-end NAS. Software raid will probably also be a thing.


By then even those will not be enough for your Facebooks, Slacks and Discords, because Chinese Silly Valley will still be too busy churning out Javascript (or the hipster language of the year) bloatfests. We don't need faster CPUs - what we need is for us coders (yes, that includes me!) to stop considering that "hardware is cheap, developer time is not". Hardware is NOT cheap when it's your users who have to afford the needless upgrade bills!

I'm all for faster video encoding boxes and sub-second Linux kernel build times, but if the main force for upgrading is being able to run the latest $SOCIALFAD whose functionality at its core is no different from '90s IRC (but with emojis!), then I'll stick to my museum-grade hardware, thanks. And don't get me started with videogames (remind me why we play videogames, pretty please? It's not for the specs, dummy!)

Licensed Pirate® since 2006, 100% Buttcoin™-free, enemy of All Things JavaScript™
Posted on 20-09-08, 14:29 in Migrating Windows 10 from VirtualBox to libvirt (revision 1)
Dinosaur

Post: #769 of 1285
Since: 10-30-18

Last post: 16 days
Last view: 2 hours
I see myself migrating away from VirtualBox in my not-so-distant future, if this ridiculous Debian-Orrible licensing/bugfixing spat continues.

It's the time to seriously research into libvirt and friends. I'm no friend of GNOME software, but on virtualization you're expected to get your hands dirty with CLI anyways.

BTW: is your W10 VM Genuine Microsoft Software™? Preserving activation on my XP VM from VPC/VMWare to VBox was anything but straightforward - to this date it still involves the use of hacked VBox BIOS dumps, for which Orrible folks don't even want to talk about (in comparison, VPC->VMWare is straight forward as long as you obey the MS licensing restrictions).

But then that's a very special case - the usual is having to call MS and get it done. Did you had to reactivate your license or whatever?

Licensed Pirate® since 2006, 100% Buttcoin™-free, enemy of All Things JavaScript™
Posted on 20-09-09, 18:12 in Computer Hardware News (revision 1)
Dinosaur

Post: #770 of 1285
Since: 10-30-18

Last post: 16 days
Last view: 2 hours
So people prefer the "disposable, portless, soldered everything" future to the "we're in total control of our lives" future because... smaller is better?

And people ask me why I wonder computers are doomed :/
I suggest to stop calling such crippled simple devices "computers" and instead refer to them as appliances, because that's what they are at the end: turn it on, press buttons, get results.

No, I don't want a freakin' IBM mainframe on my room (although I wouldn't mind re-purposing one as a storage locker :D). But I also don't want a computer I can fit inside a book either! Not as my main system / daily driver, at least - I can have fun with appliances, but I can't depend solely on these for being a fully functional person. I want ports. I want slots. I want expandability. I want to be able to get my shit working again by replacing whatever part it failed instead of "ooooh, time to buy a new appliance"!

But then, that's just me. If you're happy with your appliances and are more productive with them, more power to you.
Just leave our big fat ports-a-galore boxes alone and we good.

Licensed Pirate® since 2006, 100% Buttcoin™-free, enemy of All Things JavaScript™
Posted on 20-09-13, 02:00 in (Mis)adventures on Debian ((old)stable|testing|aghmyballs) (revision 2)
Dinosaur

Post: #771 of 1285
Since: 10-30-18

Last post: 16 days
Last view: 2 hours
In one of those very, VERY rare times when I have unrestricted/unmetered access to the Internet (not thanks to CANTV!), I managed to install nearly a year of pending updates across my active Debian fleet. I did a full "apt-get -t buster-backports dist-upgrade" to nicely bridge the gap between ol' stale Stable and the party zone at Testing. Among other things, this meant all my machines got the latest kernel on Debian, which unsurprisingly is the already-EOL 5.7 branch.

No incidents to report... except for THIS.
Those came from my oldest active Buster setup, the IBM TV box with its patched-up ATi RV630 (HD2600 in marketingspeak). Apparently the new kernel hates my card, but only in X11 - KMS framebuffer console is spotless. The wallpaper is fine, but everything else renders with that kind of corruption! Even weirder: if you look at the last screenshot, videos and 3D workloads are corruption-free - it's only things like window borders, widgets, icons and everything with which you interact with your PC! (I'm quite surprised the corruption could be captured on an ordinary screenshot - no blurrycams needed!) Needless to say, this is no bueno - this is MALO, MUCHO MUY MALO!

Things used to work fine with kernel 5.3, and indeed if I reboot to that kernel, display works flawlessly. I tried installing every linux-image release on Backports, and every since kernel starting with 5.4 exhibits the very same symptoms: consoles and 3D/videos are fine, desktop is corrupted, so it's clearly not a Mesa or Xorg bug. However, there is another workaround which I found the only one to bring some sanity back: disable KMS. With no modesetting (radeon.modeset=0), X11 is once again corruption-free, but as you already know, this kills hi-rez consoles, bootsplashes, and in general modern kernels and GPU drivers don't really like it that much (I even got the DRM bits yelling at me at the kernel log that "UMS is not compiled-in on this module" or something down these lines!). Of course this could imply that the next Debian release (Bullseye) would be a dud for this specific configuration, which would be a shame as it feels quite speedy with a half-decent GPU and plenty of RAM!

Unfortunately, while I've got some good material for a bugreport, why even bother filing a report!? To keep it languishing aside the other couple reports I've filed on kernel bugs that hit me with ZERO ACKNOWLEDGEMENT AT ALL, not even basic triaging?! Yeah, now that upstream is seriously thinking about killing AGP support I guess the only answer I would get nowadays is a "sucks to be you", not even "have this virtual money and go buy a virtual video card that won't even fit on your dinosaur" :/

Licensed Pirate® since 2006, 100% Buttcoin™-free, enemy of All Things JavaScript™
Posted on 20-09-13, 04:06 in Computer Hardware News (revision 1)
Dinosaur

Post: #772 of 1285
Since: 10-30-18

Last post: 16 days
Last view: 2 hours
Didn't noticed this as I'm finally catching up with the final batch of Win7 updates (I haven't even booted to W7 at home THIS YEAR): as a final "fuck you and thanks for all the fish", MS broke your wallpaper on the final update, leaving you with a solid black background (the punishment pirates get when their Windows is deemed "not genuine" by the system)... but only under specific conditions (you have to use the "stretch" mode).

Originally there was no fix, except to downgrade to Wintendo Diez rollback the patch, but I guess that after the hordes of angry tweets, MS had to push an exceptional ~50MB patch to undo their silly fuckup. You won't be getting this one from WU, though - you have to source it from the Catalog instead. Install, reboot, enjoy XP4LYFE best Windows~

Sidenote: Windows releases past XP really dislike being booted only a couple time a year. They get all wonky (missing icons, activation failures, excessive slowness), but all you need to restore things to normal is another couple of reboots.

Licensed Pirate® since 2006, 100% Buttcoin™-free, enemy of All Things JavaScript™
Posted on 20-09-13, 14:54 in I've been called a SNES-hating Genesis fanboy. (revision 1)
Dinosaur

Post: #773 of 1285
Since: 10-30-18

Last post: 16 days
Last view: 2 hours
I don't know what's worse: next-gen specs whores (where "how many SSD TFLOPS can drive my orgasm?" is the order of the day instead of "will we have games other than bloody FPSes or EA sportsball games?"), or retro fanboys (where the Sega-vs-Nintendo console war is still alive and bitter than ever for no good reason at all, worse than a ex-Yugoslavian republic)

I feel most of those people forgot why we play videogames in first place :/

Licensed Pirate® since 2006, 100% Buttcoin™-free, enemy of All Things JavaScript™
Posted on 20-09-13, 15:00 in Computer Hardware News (revision 1)
Dinosaur

Post: #774 of 1285
Since: 10-30-18

Last post: 16 days
Last view: 2 hours
Apparently nVidia's checkbook is the only one willing to go wide open for that ARM buy: Masayoshi Son is set to sell ARM to greedy greens for $40 BILLION. Turns out that if this comes through, Son would end being a lucky moron as he would be making a $8B profit on the sale. Too bad he will keep blowing those monies on hookers and blackjack unicorns :/

(Does anyone have a non-paywalled link? All sources I can find are the WSJ and Financial Times, which as expected want your money and personal info to read beyond the headline)

In the meanwhile, this is gonna become yet another regulatory disaster (divestment hell time, or nVidia will be very busy buying lawmakers) - Apple, Qualcomm, Google, and Samsung have very good reasons to be against the sale. Where is that "Save ARM" consortium / crowdfunding initiative?!

Licensed Pirate® since 2006, 100% Buttcoin™-free, enemy of All Things JavaScript™
Posted on 20-09-14, 14:16 in Computer Hardware News (revision 2)
Dinosaur

Post: #775 of 1285
Since: 10-30-18

Last post: 16 days
Last view: 2 hours
Posted by Screwtape
Is NVIDIA's announcement enough?

I believe Apple, Qualcomm and maybe Samsung have perpetual licences, so it doesn't matter what NVIDIA does, they can noodle around and do their own thing. However, one of the big advantages of ARM is its popularity - if NVIDIA screws up and the low end of the market moves over to RISC-V or OpenSPARC or something, those companies might want to shift over as well.

The official NVDA press release

I see too much "AI" buzzword bingo BS, but aside of that... welp, it's official. Masayoshi Son, thanks for ruining computers and fucking the UK pride!. Yeah, gotta grab some popcorn to watch the regulatory shitshow unfold, particularly at the EU and China. Fun fact, a lot of people believe nVidia is a Chinese corporation, a-la Huawei. But nope, it's American - its founder is half-Taiwanese (y'know, from the original China, fuck PRC's "One China" BS), but yeah, people can't even read these days.

> RISC-V
And now I can see the huge smile at Western Digital SoC designers when they told Marvell that they were going all-in with this newfangled RISC-V thing. Now I guess there must be a huge party going up there celebrating "the death of ARM" or something :D
Can't wait for those $50 RISC-V Android phones if nVidia screws it up this acquisition.

There IS a possitive side of all this mess: the faster nVidia kills Mali, the better. Or maybe will they finally have a half-decent binary blob driver, at least?

Licensed Pirate® since 2006, 100% Buttcoin™-free, enemy of All Things JavaScript™
Posted on 20-09-15, 13:50 in Board feature requests/suggestions (revision 1)
Dinosaur

Post: #776 of 1285
Since: 10-30-18

Last post: 16 days
Last view: 2 hours
So... MySQL died yesterday, right?

Time to migrate to PostgreSQL, yo.

Licensed Pirate® since 2006, 100% Buttcoin™-free, enemy of All Things JavaScript™
Posted on 20-09-15, 14:57 in Board feature requests/suggestions
Dinosaur

Post: #777 of 1285
Since: 10-30-18

Last post: 16 days
Last view: 2 hours
At least I like that the page footer (including the "ad") doesn't break even when the rest of the board can't go beyond a dead database.

Licensed Pirate® since 2006, 100% Buttcoin™-free, enemy of All Things JavaScript™
Posted on 20-09-15, 18:49 in Board feature requests/suggestions
Dinosaur

Post: #778 of 1285
Since: 10-30-18

Last post: 16 days
Last view: 2 hours
Posted by funkyass
why is mysql dying anyway?


It has been dying since Oracle bought them, I heard.

Now seriously, who hasn't ever visited a forum/message board which hasn't been down at least once due to some MySQL being dead? Most of the times when you get the luxury of some server leaking exception details, it's a crashed table of some sort because some process exited abnormally.
(Incidentally, another board on my bookmarks is also dead right now too due to MySQL, but this time all I get is a unlovely "Error loading character set utf8mb4" error instead of the usual crashed table shenanigans)

Licensed Pirate® since 2006, 100% Buttcoin™-free, enemy of All Things JavaScript™
Posted on 20-09-15, 23:42 in (Mis)adventures on Debian ((old)stable|testing|aghmyballs) (revision 1)
Dinosaur

Post: #779 of 1285
Since: 10-30-18

Last post: 16 days
Last view: 2 hours
Posted by kode54
Posted by wertigon
Posted by tomman

No incidents to report... except for THIS.


Oh, wow... Time to get a few used RX 550 cards?

They make RX 550 cards in AGP form factor?


Most likely this also hits the native PCIe versions of those R600 cards, and who knows what else (as ATi/AMD maintainers are focused on recent hardware nowadays). Disabling KMS fixes desktop, but kills framebuffer consoles (and I'm unsure if UVD/VDPAU/VA-API still works without KMS)

Another feature killed by recent updates: being able to use dumbphones with ModemManager: my ol' trusty (and now headless) RAZR V9x no longer gets recognized by ModemManager - not that it ever worked with it: MM always died trying to use my 3G connection (it sent some AT command related to APNs, the phone returned some empty response, MM went ballistic and segfaulted - the same phone works just fine with pon/poff and nearly every other modem dialer), SMS support was spotty (with ModemManager-GUI I wasn't even able to READ messages until a 2018 update or so, and even then it would WIPE the phone storage every time I pulled the messages out of the phone, unlike most Windows tools and Wammu). Ah well, I guess only dinosaurs use dumbphones, even if you're forced to interact with them through a PC because noone repairs broken flex cables on flip phones anymore. At least I learned how to use USSD codes via AT commands (AT+CUSD=1,"*144#") since that's the only way you can check your prepaid balance on Movistar phones over here, and even THAT was broken on ModemManager too!

The Broadcom NIC on my Dell still causes the kernel to complain, but iommu=soft still keeps things under control - that's yet another bug noone cares about fixing anytime soon (on that related Ubuntu bugreport, people hasn't had the same luck as me with this workaround, but it's fine - who uses wired NICs nowadays?).

I had finally to bite another bullet and install VirtualBox from official upstream .DEBs since the packages on Sid are too new for Buster, and the old Sid version I had installed now has an incompatible kernel module which breaks your setup on anything past 5.4 - great, more bloody red Orrible® branding on my desktop. Guess next time I have unmetered access I'll play a bit with KVM/QEMU and friends, as VirtualBox is slowly earning a place into my shitlist :/

Licensed Pirate® since 2006, 100% Buttcoin™-free, enemy of All Things JavaScript™
Posted on 20-09-19, 16:33 in (Mis)adventures on Debian ((old)stable|testing|aghmyballs) (revision 2)
Dinosaur

Post: #780 of 1285
Since: 10-30-18

Last post: 16 days
Last view: 2 hours
While my date of "ditching VirtualBox and other Orrible® products for good" is still far, FAR away, I still get to deal with VBox on computers that are not mine (that, and sadly VBox is the only user-friendly, cross-platform, free/open source VM product out there - VMWare is mostly dead thanks to Dell, Hy-Perv is a Windows-only mess, Parallels is the best Mac users can get... and even they have to pay for it, and KVM/QEMU have a long road to march towards dumbproof despite being already production quality).

This time, yet another P2V situation: $NEW_CUSTOMER has their current $PRODUCT running on a long EOL'd Ubuntu 10.04 LTS HDD that came from a unknown computer, then ended at a not-so-shabby ThinkServer TS140 (a complete waste given that we're still dealing with no UEFI support, a 2.6.32 kernel, and a non-server Ubuntu setup pressed into production by $SOME_MORONS). The product in question is another ancient, unsupported LAMP-based payroll system they're still using yet they're about to migrate to the solution sold by $FORMER_BOSS, and I've been contracted to P2V this crap so they can actually use both applications during the transition without having to reboot several times a day.

Long short story: P2V'ing ancient Ubuntu is a piece of cake these days if you still choose to stick to VirtualBox.

- Install Orrible® upstream .DEBs, as we're on Buster which means no official Debian packages, ever. You also want to RDP into this thing (because LOLubuntu and its unlocked desktops by default!), so make sure to install the "Oracle VM VirtualBox Extension Pack" as you will need that one for MS-compatible RDP (VNC support is only present on FOSS/community/distro builds, not on upstream binaries), USB2/3, PXE for some emulated NICs, etc.

- On another PC running VBox, prepare the source HDD image: clone the original HDD to another disk (we MUST NOT TOUCH the original production disk until we're already on the production VM!, plus backups), clear enough space on the clone to shrink the partition (those guys loved plaintext .SQL dumps for everything, so I was able to compress ~35GiB of "backups" to ~890MB in .tar.xz, bringing down the target HDD image from 250GB to 50GB with plenty of space to spare).

- Attach the clone HDD to VBox (create a physical VMDK, don't forget to change permissions to 666 on /dev/sdYourCloneDisk, umount any partitions here prior to continuing!), create a new Ubuntu 64-bit VM (gib enough RAM, add a new blank VDI disk large enough for the target, we want bridged networking, VBox 6.1 now defaults to VMWare SVGA video which makes our life easier should we want to not install guest extensions), connect our physical VMDK to it, boot to test.

- Does the VM boot OK? Test ACPI shutdown, etc. Boot the VM again, but this time use your favorite Clonezilla ISO to start it, clone the physical HDD to our final, blank VDI. Move/enlarge partitions as needed (if you wish/have OCD and can't stand unpartitioned space). Shut down the VM, detach the physical VMDK, test boot again from the VDI. We're now ready for the final and most glorious step...

- ...which is exporting the VM to the actual production server! To save some time, we will export the OVA appliance directly over NFS (so don't forget to have a speedy CPU, gigabit LAN if possible, and nfs-kernel-server installed on both hosts!). This is the most time-consuming part of the process, albeit user interaction is minimal (in my case, the final OVA shrinked the VM further down to ~2.7GiB, hooray for cheapass LAMP bloat!). Get intimate with VBoxManage, as we're now on headless CLI-land!

- Import the VM (VBoxManage import yourappliance.ova; test with --dry-run first!), fix network settings (VBoxManage modifyvm yourVM --nic1 bridged --bridgeadapter1 eth0), enable RDP (VBoxManage modifyvm --vrde on, set authentication if desired), now start the VM (VBoxManage startvm yourVM --type headless), check network connectivity, RDP into the VM and fix network settings (IP address, etc.), test shutdown (VBoxManage controlvm yourVM acpipowerbutton, then wait until the VBox host process exits). Test everything twice! You're done... almost!

- To seal the deal, we now need the VM to come up on every server boot, and to shutdown cleanly at every server shutdown. This is where things have always been a bit of a sore thumb on VBox, particularly with Debian. If you haven't already noticed, VBox ships with a vbox-autostart initscript, the setup procedure is not straightforward and the whole mechanism is as fragile as a house of cards (VMs may fail to start or shutdown cleanly, getting permissions right is tricky, system updates may break everything, and it never worked properly at Debian, whose VBox packages never shipped the vbox-autostart script at all!). Luckily, if you're not a Troo UNIX® Way graybeard a Poetteringware hateboner allergic to systemd, getting VBox autostart with it working is surprisingly easy AND reliable - all you need is a custom unit file, run a couple commands, and you're set for success. No permissions fuckery needed! Try it. Try bringing up and stopping VM. Try rebooting your server and watch. Now We Are Playing with (Virtual) Power! Call it done, crack open a cola/beer can, go home and enjoy your lenghty decontamination shower. You've earned it.

- If everything went according to plan, before leaving, ask the local slaves to access both the old and new applications. They should notice no difference at all, they aren't even aware that their old crusty crapp is now running on fake hardware :P

- The old production Suckgate HDD? (because it had to be a Suckgate 7200.12 HDD with >60K hours on the clock and broken, unsafe firmware, that's why!) Backup that one for a final time, hand over that backup to the resident IT drone, tell customer you've reinstalled it but left it unplugged, make it magically vanish.

Licensed Pirate® since 2006, 100% Buttcoin™-free, enemy of All Things JavaScript™
Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
    Main » tomman » List of posts
    Yes, it's an ad.