ZFS bread

>testing on server
Is it fine? I'm considering either debian stable or ubuntu server for a small fileserver at home. Also heard ZFSoL has been jumping between good and bad still.

Attached: increasing degrees of sweating.jpg (480x1440, 139K)

Friendly reminder Hans Zfs murdered his wife.

I've had a few issues where I had to fiddle with dkms, spl, kernel, zfs packages, their versions, etc (configuration stays the same, so it's not really that risky), but it's not too bad. I just like having it up to date. And since I do backups of it every now and again and it's not much more than a NAS (and XfR router), I won't lose much if anything goes wrong.

Where is my fucking block pointer rewrite?

Bloated TRashy FileSystem. It sucks XXL Chance Flared.

>Hello valued customer there will be maintenance at 1AM to update backend storage
Every fucking time.
Is there a way to have ZoL update without fucking everything up?

zfs is amazing

Using ZFS on Proxmox for VMs and containers was braindead easy to set up and has been working flawlessly for 10 months on my dedicated server.

if zfs is deprecated and btrfs is buggy (and retardedly named), why would anyone use this shit

>Lets talk about the best filesystem ever made
bcachefs?

>can't even do union mounts
use v9fs instead, or better yet just install plan9

fat32 forever

>ZFS is deprecated
U wot

I use it on my laptop and my company uses it on all our storage servers. It's nice.

Does Hammer2 have the ability to incerase performance via caching on ram or in ssd like zfs does with arc and l2rc?

>murderfs

I use ext4 cuz I'm chad

Why use zfs over btrfs?

I have two freenas boxes. I've fucked around with lots of configurations before I just installed freenas, I'm pretty happy so far. Support community is absolute fucking cancer and worse than Arch community. BTRFS and LVM was fun to mess around with and pretty fast too.

how is it better than plain ext4 (for desktops)?

muh bitrot

No. I use HAMMER2.

see

why? now that ZRAID has been deprecated by LVM there's no reason to use this license-bloated piece of ancient trash.

>buggy
Claim unsubstantiated

>abandoned
huh

>le sucks xddddd
You gotta be 18 to be on Jow Forums

What are you Jow Forumsentoomen using, FreeBSD, ZFSoL, Freenas or perhaps NetBSD?
I have a lot of hardware and I'm not against running a specialized install for my storage server.

ZFS on Linux. more specifically Debian stretch. I have no complaints about it, it just werks

Attached: 1339550542728.jpg (480x451, 57K)

I have serious reservations about basically how relatively poorly all the interesting features in ZFS perform.

But btrfs isn't really any better, it's still buggy and many of the interesting features aren't implemented either with glacially slow progress towards them and it not actually looking too good on the performance end in the process.

It sounds like one of the best, but it's more "outlined" than "made".

The interesting stuff like configurable erasure coded (RAID-type) storage isn't even implemented yet, I think.

Does it allow for file tagging like a DBFS? No? Then fuck off you cunt.

> file tagging
Not sure why you'd ever want that in the filesystem rather than in an application running on the filesystem.

Nothing beats native implementation. And the programs there are all suck ass.

Switched to Btrfs last year. It's reliable enough now and seems to work better on RAMlet devices.

> Nothing beats native implementation.
That means basically nothing. You'd still do some kind of xattr thing maybe with its own permissions (can you read/modify/delete tags?) and crap and it wouldn't really help much at all that it's in the filesystem.

It'd probably just commit us to hash method xy up to length z forever for fear of breaking that implementation in the file manager on Windows and OSX and that popular python lib... or whatever.

> And the programs there are all suck ass.
Make a good one if it bothers you?

I don't want something that just werks, I want the maximum autism best choice available, since I'm not an enterprise and just doing this for fun.
I want to try something new, so running it on a minimal Linux distro would be boring. And I've heard bad things about FreeBSD's stability in general.

Does anything do it well?

so... you deliberately want something that's broken and that doesn't work so you'll have something to tinker with?

wouldn't you rather set up something that works well so that after you've learned about it and tinkered with it you can feel satisfied, enjoy your new fileserver or whatever you build, and then go on to find some new project to learn and tinker with, while the old one just keeps ticking away making your life better without additional fuss?

Why are Debian users always such cunts?
No, I run multiple Linux boxes, I just want to try some new technology. Sometimes you have to expand your horizons and learn and work on something new. Additional fuss is good for your brain.

I don't know about FreeBSD in general, but my FreeNAS install has been rock solid for a year now, averaging 2 months uptime, only going down for updates or because I needed to work on the box. ZFS has a shitload of things you can configure and tune, it just comes with defaults that are within 15% of optimal for almost every situation.

I was kinda assuming that you wanted the tinkering and additional fuss to, you know, eventually wind up with something useful. If you just want the fuss for the fuss's sake, why do you need our advice? go take apart your vacuum cleaner and then put it back together or something.

Exactly, why not both? Hence my question.
Go update some LTS packages for CVEs or something.
Do you run ECC RAM? How light is freenas on system resources?

Why stop at one choice then? Try them all out and see what works best for you.

Why would you run ZFS on the desktop, talk about overkill and waste of resources.

>because of the license
This isn't a valid concern for most users. ZFS is what BTRFS wants to be, except it's already done AND has years of tooling built around it, testing done on it, and ports to other platforms.
When picking a file system, I want the technically best one, not the legally "best" one.

>but it doesn't ship in-kernel with my distro
Either build your own kernel or use a distribution that does include it. Like one of the BSDs or Solaris systems. It's not like your choices are limited.
Meanwhile with BTRFS I hope you want to use GNU/Linux exclusively. Even Windows users have ZFS.

How is it a waste in any regard? It's not like it uses more than other file systems by default, but it does offer better functionality and some efficiency features too if you want them. Who wouldn't benefit from sendable snapshots?

I'm pretty sure btrfs didn't want to be as slow as zfs, or as unable to add / remove drives from raidz or as high latency on raidz or any of these things.

But yea, they geelnerally havent managed to do better either.

>unable to add / remove drives from raidz
What implementation are you using? I'm pretty sure OpenZFS handles this as expected.

Ceph - scales out and gets better the more hosts/disks you give it, multiple failure domains, can actually expand with any size disk

Only since 2017

Attached: file.png (609x256, 25K)

It wasn't merged yet, was it? It certainly wasn't in 2018.

>I've heard bad things about FreeBSD's stability in general
I've had the opposite experience. I ran it on an x64 machine for multiple years uninterrupted outside of voluntarily doing (built from source) release updates. I also ran it on my PS3 because. That is noteworthy because I used to compile software for PPC64 on it which is abusive on that hardware. I didn't have any issues as long as I didn't OOM (PS3 has 253 usable 2MB for HV).

That being said I've since migrated to Openindiana on machines I use daily and have been enjoying it.

It depends on where you mean. I beleive it's in OpenZFS but no distro I know is using it. FreeBSD 12 is supposed to ship with it, I don't know if one of the Solaris systems already includes it or is probably also still testing.
I have to imagine they're really careful about introducing FS features into mainline.

Attached: PSBSD.png (571x495, 51K)

It's a nice filesystem, but describing it as "just werks" is more than a stretch.

It is complicated AND buggy as fuck at this point. With a pretty much overly unintuitive CLI and poor presentation of information.

>It depends on where you mean.
I'm saying it was announced, but never merged into the production ready stable code base.

You "believe" this is wrong? Okay, where is it? I can't see it, and when I checked just now secondary sauces like:
github.com/zfsonlinux/zfs/issues/8437

Indicate that the respective code development has stalled and that people wish someone else would pick it up and get it done.

Obviously it is not at all done, as far as I can tell, even if it was announced in 2017 and other software has this feature.

FreeBSD had an experimental patch in April 2018 reviews.freebsd.org/D15124
The design doc for other operating systems to implement is also linked, so an incarnation exists and the source is included there. This probably has everything necessary to find the branch you're looking for upstream.

But remember, they have these testing periods for a reason. Would you really want your volume manager to be unstable when adding or removing a volume? I'd wait for FreeBSD to determine stability and additional patches first, but that's me.

By "merged" into I mean the usual meaning of into some mainline release ZFS.

Not a "pre-alpha" prototype code in a developer repository / patch diff somewhere. The latter are features you usually aren't supposed to use unless desperate or a developer,

Basically, ZFS does not have RAID expansion yet, in spite of it being announced in 2017.

I guess email Matthew and ask him what branch he's pushing to now. It makes sense that the code would live in some OpenZFS repo somewhere not the FreeBSD one after moving out of alpha. Alternatively email Alan. If it's a feature to be in 12, I bet he'd know the upstream status.
Not really sure to be honest. I haven't needed that feature myself. I always use 4 disks and I upgrade them in pairs of 2. So I haven't been following it.

> after moving out of alpha
See, I don't think this actually happened.

That's why I'm saying ZFS (and OpenZFS) don't have this feature.

Fair enough. I don't know enough about what's going on here to say for certain. I have a hard time believing development for a highly requested feature suddenly stopped, and at least for the FreeBSD "3 months after it's done" cadence I expect, they're still in a sensible release timeline for a 12.X feature. (meaning I would expect that to land sometime this year if it is actually still in development).
They have code, design docs, specs, discussions, talks, etc. The components are all there to make a stable patch. I imagine they're doing just that. Which probably takes an eternity to test for something like ZFS.
>multiple configs on multiple operating systems
Fuck that.

Is ECC required? I suspect yes, but I've seen some retards say ZFS works fine without ECC RAM.

What a stupid way to phrase that question.

u don't need ECC why would you think that?

I suspect no, and I've seen some retards say ZFS doesn't work fine without ECC RAM.

>ZFS
absolute trash
>BTRFS
absolute trash

>mfw in the process of dropping Btrfs right now

Attached: 1501709332226.png (778x1018, 591K)