Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Linus: Don't Use ZFS

realworldtech.com

Post reply on HN
Print view View original

341–350 of 571 posts

Re: Linus: Don't Use ZFS

#341

Post by jupp0r »

zippergz wrote: post #201
m0zg wrote: post #46

Earlier quoted context omitted.

Pretty much my view as well. If Linux becomes incompatible with ZFS in any way, I'll switch to FreeBSD. That said, after the Oracle Java debacle, I can see why Linus would not be receptive towards merging ZFS into the kernel. I just wish he argued the point on legal issues alone instead of making up stories about non-existent technical flaws in ZFS. The whole thing is basically a work of art. Oracle should consider G…

I have been using FreeBSD for its better ZFS support for years, and it's great. Highly recommended.

FreeBSD has been using the ZFS on Linux code for a while.

Re: Linus: Don't Use ZFS

#342

Post by nickik »

m4rtink wrote: post #258
reacharavindh wrote: post #10

As a heavy user of ZFS and Linux, what else is there that even comes close to what ZFS offers? I want cheap and reliable snapshots, export & import of file systems like ZFS datasets, simple compression, caching facilities(like SLOG and ARC) and decent performance.

XFS on LVM thin pool LV should give you a very robust fs, cheap CoW snapshots, multi device support. If you want, you can make the thin pool be on RAID via LVM RAID under the thin pool. For import export, IIRC XFS has support for it and you can dump/import LV snapshots to get atomicity. For caching there is LVM cache, should be again possible to combine with thinpool & RAID. Or you can use it separately for normal LV…

Call me when somebody like a major cloud provider has used this system to drive millions of hard-drives. I'm not gone patch my data security together like that.

There is difference between 'all these tools have been used in production' and 'this is an integrated tool that has been used for 15+ years in the biggest storage installations in the world'.

Re: Linus: Don't Use ZFS

#343

Post by nickik »

bayindirh wrote: post #29
reacharavindh wrote: post #10

As a heavy user of ZFS and Linux, what else is there that even comes close to what ZFS offers? I want cheap and reliable snapshots, export & import of file systems like ZFS datasets, simple compression, caching facilities(like SLOG and ARC) and decent performance.

Honestly asking, how Btrfs compares to ZFS? There's also Lustre but it's a different beast altogether for a different scenario.

Btrfs has eat my data, and once that happens I will never, ever, ever, literally ever go back to that system. Its unacceptable to me that a system eats data specially after multiple rounds of 'its stable now'.

But in the end it always turns out that only if you 'use' it correctly it is actually not gone eat your data.

I used ZFS for far longer and had far fewer issues.

Re: Linus: Don't Use ZFS

#344

Post by nickik »

montjoy wrote: post #83
reacharavindh wrote: post #10

As a heavy user of ZFS and Linux, what else is there that even comes close to what ZFS offers? I want cheap and reliable snapshots, export & import of file systems like ZFS datasets, simple compression, caching facilities(like SLOG and ARC) and decent performance.

Hardware RAID controllers can do most if not all of these things.

Pay more for less safety and put all your data into the hands of the guy who wrote the firmware for that thing. I'm sure that software is well maintained open source code.

Re: Linus: Don't Use ZFS

#345

Post by mschuster91 »

ac29 wrote: post #295
liuliu wrote: post #166

Earlier quoted context omitted.

bcachefs should be heavily supported, it doesn't get nearly enough for what it supposes to do: https://www.patreon.com/bcachefs

Many of the advanced features aren't implemented yet though, like compression, encryption, snapshots, RAID5/6....

why would you want to embed raid5/6 in the filesystem layer? Linux has battle-tested mdraid for this, I'm not going to trust a new filesystem's own implementation over it.

Same for encryption, there are already existing crypto layers both on the block and filesystem (as an overlay) level.

Re: Linus: Don't Use ZFS

#346

Post by nickik »

gigatexal wrote: post #5

He mentioned that he didn’t think it was being maintained. It’s more or less been formed no? Has Linus not seen the work that the OpenZFS folks are doing? ZFS is amazing and I would soon go to a BSD flavor with a fun set of user land utilities than give it up.

Its pretty clear that Linus simply doesn't have a clue about ZFS. And he just exposed himself as somebody that repeats stuff he read in some linux forum or something.

There is no way, after any technical evaluation by himself he would come to those conclusions.

Re: Linus: Don't Use ZFS

#347

Post by Youden »

brianpgordon wrote: post #158
Youden wrote: post #56

Earlier quoted context omitted.

I've tried btrfs without much luck. btrfs still has a write hole for RAID5/6 (the kind I primarily use) [0] and has since at least 2012. For a filesystem to have a bug leading to dataloss unpatched for over 8 years is just plain unacceptable. I've also had issues even without RAID, particularly after power outages. Not minor issues but "your filesystem is gone now, sorry" issues. [0]: https://btrfs.wiki.kernel.org/in…

My home NAS runs btrfs in RAID 5. The key is to use software RAID / LVM to present a single block device to btrfs. That way you never use btrfs's screwed-up RAID 5/6 implementation.

If you use LVM/mdadm for RAID, it's not possible for btrfs to correct checksum mismatches (i.e. protect against bitrot).

Re: Linus: Don't Use ZFS

#348

Post by Godel_unicode »

metaphor wrote: post #288
berti wrote: post #230

Earlier quoted context omitted.

Keep in mind that RAID5 isn’t feasible with multi-TB disks (the probability of failed blocks when rebuilding the array is far too high). That said, RAID6 also suffers the same write-hole problem with Btrfs. Personally I choose RAIDZ2 instead.

To be sure, it's entirely feasible, just not prudent with today's typical disk capacities.

Feasible is different than possible, and carries a strong connotation of being suitable/able to be done successfully. Many things are possible, many of those things are not feasible.

Re: Linus: Don't Use ZFS

#349

Post by znpy »

bayindirh wrote: post #57
znpy wrote: post #48

Earlier quoted context omitted.

people probably will, in a few years. rotational disks are getting cheaper and cheaper, 10TB disks in two years might cost as little as 2TB disks today (i got a 2TB disk for like 50€ off Amazon).

> people probably will, in a few years. Yes, but without the array as you stated. We have 300+ 10TB disks at our datacenter today and, ZFS is relevant at this disk count, I/O and client load. Running ZFS at small scale is raising a cow at home for a bucket of raw milk. It's more of a fun curiosity rather than a production level operation. I'd run LVM or md or something similar at home instead of a full blown ZFS setu…

> Running ZFS at small scale is raising a cow at home for a bucket of raw milk. It's more of a fun curiosity rather than a production level operation.

Well on one hand this is true, but on the other hand...

If you're running more than one disk at home, maybe you're some kind of enthusiast (homelabber?) And willing to put some effort into that. Under this scenario, the same amount of time spent learning zfs yields better results vs lvm/mdadm.

Re: Linus: Don't Use ZFS

#350

Post by Youden »

beatgammit wrote: post #314
Youden wrote: post #56

Earlier quoted context omitted.

I've tried btrfs without much luck. btrfs still has a write hole for RAID5/6 (the kind I primarily use) [0] and has since at least 2012. For a filesystem to have a bug leading to dataloss unpatched for over 8 years is just plain unacceptable. I've also had issues even without RAID, particularly after power outages. Not minor issues but "your filesystem is gone now, sorry" issues. [0]: https://btrfs.wiki.kernel.org/in…

Why use RAID5/6, RAID10 is much more safe because you drastically reduce the change of a cascading resilvering failure. Yes, you get less capacity per drive, but drives are (relatively) cheap. I thought I wanted RAID5, but after reading horror stories of drives failing when replacing a failed drive, I decided it just wasn't worth the risk. I currently run RAID1, and when I need more space, I'll double my drives and s…

I use RAID6 because it gives me highly efficient utilization of my available storage capacity while still giving me some degree of redundancy should a disk fail. My workload is also mostly sequential, so random read/write performance isn't too important to me.

If a disk fails and resilvering causes a cascading failure, I can restore from a backup.

I think you might be mistaking RAID for a backup, which is a mistake. RAID is very much not a backup or any kind of substitute for a backup. A backup ensures durability and integrity of your data by providing an independent fallback should your primary storage fail. RAID ensures availability of your data by keeping your storage online when up to N disks fail.

RAID won't protect you from an accidental "rm -Rf /", ransomware or other malware, bugs in your software or many other common causes of data loss.

I might consider RAID10 if I were running a business-critical server where availability was paramount, or where I needed decent random read/write performance but even so I'd still want a hot-failover and a comprehensively tested backup strategy.

Post reply on HN