RIP Linus Tech Tips for using cuck license filesystems...

Has nothing to do with a filesystem you retards, misconfigured the fuck out of the server and that's what caused the data loss.

that's only two disks in a vdev, did he not use raidz2+?

>He used ZFS and still lost data on power failure
Yes because it was configured to NEVER scrub the data, which is just retarded.

If properly configured it would scrub the drives every ~30 days and upon a power failure.

Which would've been just fine.

Instead it ran for 2+ years without ever having error correction run on it.

>configured to NEVER scrub the data
>Instead it ran for 2+ years without ever having error correction run on it.
haha holy fuck, why even bother with a checksumming filesystem if you never verify them?
may as well have used shit like ext4 lol

OP image is not the root of the issue, he misconfigured the fuck out of the server which caused it to never scrub the data, didn't even have offsite backup, they had a couple vdev with millions of errors that caused to not have enough parity data to resume operation without data loss.

OP is a retard, tried to meme his way just to hate on these morons, but ended up being a tool that only reads headlines or shuts his brain off when checking content.

Dilate ZTranny. BTRFS has saner defaults and is more resilient AND is not cuck licensed, maybe an unfortunate fact for cucks like you and Linus

i can't believe this clown has the gall to think he can teach anybody else about technology

I don't even use ZFS, never did, never will, doesn't make you any less of a retard, dilate.

btrfs in /home to manage data with snapshots is pure kino though

linus has more dollars than sense, maybe he should hire a sysadmin to manage his shit properly

He's a corperate whore that is just in it for the Ad money.

I don't understand these server filesystems. Is ZFS really meant to be bad?

He's being paid by Microsoft to show how shitty Linux is and to shill windows in it's place. Nothing to see here.

The overconfidence of having a couple employees including himself that a above average knowledge over these things to do it themselves, which is retarded considering the massive scope of his company.
An actual IT expert, both for networking and managing servers is long due, he most likely will still not get one, since he banks on the content he records whenever they perform any form of maintenance.

True, there's a new video spewed out every day from this channel and they're all manufactured. The days of genuine content is gone. Has been gone for years and years. They pretty much just make up random scenarios for videos like needing to buy a chromebook or "my pc isnt working!!!" whenever they get so picky that one of the thousands of items sent to them by companies isnt worthy of a video

Attached: 1643402460418.png (131x125, 20.64K)

I'm a noob when it comes to this sort of file management. Can I get a quick rundown of ZFS, Btrfs, ntfs and all the other popular options and why some are shit and others are great?

It's amazing if you rtfm

> They pretty much just make up random scenarios for videos like needing to buy a chromebook or "my pc isnt working!!!" whenever they get so picky that one of the thousands of items sent to them by companies isnt worthy of a video
it's amazing how linus then can cry about people blocking adverts when this is the shit his channel pulls every day. it's no fucking wonder people block his adverts and use sponsorblock to skip over the garbage. he's literally begging for it due to his cancerous ways.

it's one of the most featureful and solid filesystems out there
if you don't maintain it properly then it doesn't matter what you use
just find a page describing them, there isn't really a one-liner i could write here that would make any sense to you, since especially zfs/btrfs vs. ntfs is so vastly different that the only thing they share in common is they both store files

>giving your daughter myocarditis

>Stores 1PB of data locally, in a single location with no off-site backup

Why would you need 11,000 videos in 4K anyway? Create a script to delete anything over 2 years old

i learned very quickly when trying out ZFS for the first time that hdds fail in many more ways than just dying outright
if anything that's the better option
i've had more than one hdd fail in a way that involves looking like they work, writing without complaints, but then spitting back tons of read errors, these motherfuckers will fly right under a filesystems' radar unless you scrub them

so hire someone competent to monitor things, and then when something interesting happens, have him catch linus up to pretend to fix it

I'm sure their editors like old footage for B-roll clips, and being a millionaire tech youtuber without a server would just be embarrassing.
That said, being a millionaire tech youtuber who fucks up administrating his server is also embarrassing.

>60 nvmes
>zfs settings tweaked all the fuck out to overclock speed for no real reason
>2 die
>it kills his entire pool

sounds like he didnt have enough parity and probably shouldnt have used 60 nvmes for the meme

according to his pool was never scrubbed, so while two may have outright died, others had errors which were never checked for
scrubbing is basically the first thing you read about when looking up zfs maintenance, so it's obvious they didn't read shit

my desktop has six decade-old hdd's in it in btrfs raid6 and even i know to scrub these every month or so

Linus should stick to racemixing instead of providing tech tips.

>Says in the video hardware isn't at fault
>Two drives next to each other failed
That's almost always the hardware's fault. Could be a hot spot, drive vibration issues, chassis fit finish problem, backplane shorting, any combination of the above. Bit it's all hardware. Drives failing next to one another is rarely coincidence.

Well, yes, it's obviously very stupid. Really what is more stupid is that the server(s) storing all their shit apparently didn't even have some UPS connected so it could shut down safely when the power goes out.

>>Isn't that crap highly unstable in RAID5/6? Have they fixed it yet?
back in like 3.x it had some nasty corner cases that could blow up the filesystem. those are gone, there's still bugs and some gotchas, but they're irritating rather than dangerous.

>there's still bugs and some gotchas
Honestly doesn't sounds like something I want to trust my data to then. Maybe in another 5 years.

I'm not sure if ZFS and BTRFS are both equally mature and feature-heavy these days or not. Also, I know that ZFS has both BSD and now apparently even Linux based "easy" distros made to configure storage, NAS etc (ie TrueNAS does both and has ZFS), but what about BTRFS? Awhile ago Rockstor used to kinda sorta be the equivalent, but it has been apparently dormant so that's an issue if there isn't something comparable to TrueNAS. Yes I know you can install ZFS or BTRFS on any distro and configure it manually etc... yeah but it is a fuck of a lot less places for a new users to fuck up and a good starting point with something like TrueNAS or a btrfs equivalent so people can see the benefit of having those filesystems for storage over just plain ext4 or something.

seriously. everytime ive watched one of his videos, its always bland delivery of a script until he gets to his faggot sponsor pun that some npcs in the comments genuinely find funny
then its just more delivery of the script until we arrive at the end, receiving a lackluster conclusion that doesnt wrap anything up (if you happened to be interested in the topic) and then again, more sponsor shit
lots of filler to make videos stretch out to 10 mins+ to get even more ad revenue as well.
and yeah, like you said its just complete cancer with how he spams ads, spams posts for the algorithm, has clickbait titles and thumbnails, and THEN puts sponsorship ads into the video. AND advertises his merchandise website constantly throughout the video
i dont know what it is with youtube, but 99% of people that get any success with it become intolerable. and it never stops. its not like they get 10 million subs or have really high view counts and then start to make more down to earth, genuine content. it just keeps going

>Li**s T*ch T**s
Can we please stop mentioning this fuck and his shitty youtube garbage?
Fuck this shitlord and his entire fakeass bullshit channel full of bad advice and stupidity.

wait so ext4 wont lose data or power failure?

Don’t you mean
>giving your wife’s daughter myocarditis

>btrfs
He would have gotten data loss without drive failure in this case.

kys smoothbrain

>endemic
why do normalfags like to poorly attribute words into their shitty vocab?
en.wikipedia.org/wiki/Endemism

The freetards wail as they strike you.

Attached: winserv.png (225x225, 5.87K)

The same guy that was absolutely certain he should uninstall his DE is also absolutely certain about medicine. I'm sure that is sound advice.

Their problems ran so deep it didn't matter what software they were using. You just can't have a setup where 5 drives in a single server can fail without any alerts. I just have a daily email to myself for my zfs status and weekly scrubs. It's annoying that it's daily but this way if my email pipeline breaks I'll notice that I'm not getting regular emails. ZFS, BTRFS, erasure coding, hardware raid, nothing can save you from letting a critical mass of hardware fail with no alerting.

ntfs is rockstable and now that it's on linux (paragon ntfs) it's a very viable option.
the fun part is ntfs has native reverse compatibility unlike other fs which can break when you don't match the versions and accessed a new version drive on old device. you won't even have to care which version of ntfs you;re using because it just werks.

ZFS is much more mature than BTRS and much more trustworthy. Personally I won't use BTRFS and I'm hoping bcachefs will be a better linux root experience.

That being said, ZFS is also a big complicated codebase and raid of any kind is not a backup. You absolutely need backups for your critical data no matter what you're doing locally. Every file system has had some form of software bug that has eaten data, even if you're more likely to have a hardware fault that zfs/btrfs/etc protect you from than a software fault.

The problem was not hardware or software related, it was that linus was a retard and did not set up both the hardware and software properly, and after an apparent failure just wiped the disks instead of looking at the actual issue.

There is a legitimate complaint in the video: there is no way to just read the data as-is on the disk with ZFS. It's been a feature request before but the devs were super retarded about it and wanted to null out the entire block rather than ever server a single bad bit, even if the user was explicitly asking for it.

holy moly this
also nothing of value has been lost

>Two drives next to each other failed
>That's almost always the hardware's fault.

Array of 15 drives. 1 fails. 2 in 14 chance next drive to fail will be adjacent to the first just from random chance. For the end cases, just treat the array as a ring.

This. NTFS is three-decades battle-hardened, and is up there with XFS for FS hours-to-unrecoverable failure ratio. Freetards don't understand that their "opinions" about NTFS are actually their opinions about their shitty reverse-engineered NTFS drivers for their hobby OS - for obvious cope-related reasons.

should have used WinFS

WinFS was a database engine that ran on top of NTFS.

I will never understand why anyone who has the option to use EXT4 doesn't.

because that goes against the design principals of ZFS
the ability to recover unverifiable data is an anti-feature, because in the case the filesystem gets in a bad state, you are meant to restore from your backups, period

-- and if you don't have a robust backup plan, why are you fucking with ZFS? may as well use NTFS and cross your fingers

I used to babysit several bare metal NTFS arrays that veeam backed up into. They were always shit.
ReFS exists because NTFS is aids. Microsoft didn't dump millions into its development because NTFS is good.

No

>i...
No, you didn't user.
>refs
Has proven to be a failure. I've seen exactly two in the wild in nine years.

>I've seen exactly two in the wild in nine years.
Well when you're managing arrays for babies down in the TB scale that wouldn't surprise me.
The good thing about passing block level access to a ntfs guest that will inevitably implode is that the array it's running on that's invariantly running some commercialized bsd/linux will happily pass back its own snapshots to vss so even a help desk monkey can usually unfuck it.

"Don't use ZFS. It's that simple. It was always more of a buzzword than anything else, I feel, and the licensing issues just make it a non-starter for me."

>Linus was right again

Is there a reason why this guy doesn't archive finished stuff on tape and put it in a fireproof vault somewhere?

Attached: tape drive.jpg (706x463, 65.14K)

doesn't look cool on camera

Get back to me when this dick actually dies

What the fuck is the point of this filesystem if you have to manually configure it to scrub to data loss on drive failures?

you just setup a cron job or equivalent to do it once and that's it

But why?
It's already running. Why can't it just scrub on its own?

this isn't a consumer filesystem, user, enterprise things like this are not expected to do things you don't tell it to do

So it exists to ensure jobs for people.

what do you want me to say, user? you think enterprise storage is being gatekept away from normal people who just want to buy a box and put files on it?
they sell things like that already, such as synology, and yes, those will scrub the disks on a pre-configured schedule, as well as a bunch of other basic maintenance tasks, they also provide support for them if you need help, this is the kind of thing that is more suited for linus
going DIY with enterprise shit is not for the faint of heart

Attached: 2022-01-30-175230_876x500_scrot.png (876x500, 156.33K)

fucking retards
Linus installed CentOS, an OS with no native ZFS support, and whoever packaged ZFS for CentOS was retarded enough to disable automatic scrubbing

On FreeBSD and Solaris, you need zero config since this is the default.

How many times has he had a file server explode now? 2/3 times? He should really get a professional to set these up.

Is pic related accurate?

Attached: bsd.png (1151x959, 655.55K)

>giving your daughter pericarditis and cardiac arrhythmia instead of myocarditis

He is retarded running that many drives in a vdev... Fucking nut.

it will corrupt it.
ups is required.
or pray that fsck /dev/almostgone1 fixes it.

sweden no

Attached: Screenshot 2022-01-29 134633.png (577x211, 28.78K)

keeeeek so not only a youtube tard but tard for real.

>ima techno rockaaaaaaa big fat rockaaaaaaaa techno rockaaaaaa
must have been that mdma he took that one time

I wonder if any of the people he has built storage servers for are feeling a little concerned about their storage servers right now.

No it won't retard

I enjoy his hypocrisy when he says he'll never say nigger right after saying faggot on camera.

>keeps larping
No, you didn't user.

You don't on distros with good support for zfs, I turned my old gaymen PC into a storage server with zfs on ubuntu and installing zfs automatically added a cronjob to scrub every month.
But regardless, anyone who spends more than 10 minutes reading up on zfs would know that regular scrubbing and error reporting is important and would check if it's enabled.

Attached: fyad.png (500x500, 37.82K)