appzer0, 10 months ago

@nixCraft Yes, #ZFS is so stable and reliable. I use #znapzend to automatically send snapshots to my backup servers. https://www.znapzend.org/

Dreit, 10 months ago

@nixCraft Reminds me of guy which copied terabytes of data on btrfs years ago. At that time it was already considered quite stable FS. For some reason he opened dmesg at same time and let it running.
Suddenly one or two copy errors appeared. When he asked btrfs about any problems, all checksums were perfectly fine, but after running script to compare physical data he found inconsistencies.

spezivernichter, 11 months ago

@nixCraft
could you elaborate a bit on that? so far happy user of btrfs, but this doesn't mean there won't be trouble in future. what risks do you see with btrfs (or xfs) ?

phaysis, 11 months ago

@nixCraft
Personally, I'm wondering how distros (Ubuntu, for example) were able to reconcile the Linux GPL and the Oracle CDDL licenses which were (famously) mutually incompatible. Is it a case where enough users said, "We want ZFS in our Ubuntu" and the lawyers figured out a loophole?

marting, 11 months ago

@nixCraft #opensuse with an external disk, #btrfs and snapshots for the backup, not one issue in almost 10 years now.

amri, 11 months ago

@nixCraft mine is XFS for DB and BTRFS for the OS 😋

#suse #sles #opensuse #hana

tmihai20, 11 months ago

@nixCraft with enough redundancy most recent filesystems can be reliable; being an active user of OpenMediaVault, ZFS needs a plugin on most common systems (plugin that comes in the form of a kernel module and is sometimes maintained only for the latest version of OMV and kernel); also ZFS needs a little more RAM

schamschula, 11 months ago

@nixCraft I‘be been relying on a #zfs pool on my #freebsd server for seven years. Over those years I’ve had to swap two drives. After resilvering everything was quickly back to normal. However, there are additional on- and offline backups as well.

TurtleCrazy, 11 months ago

@nixCraft

Definitly. Nothing more than:

zfs send -c -i zroot/dataset@T0 zroot/dataset@T1 | ssh user@elsewhere zfs recv -Fdu zfs/backup

Rolling these back or even restoring a full dataset a new machine make things (ugrades!) more confortable.

The main issue is not only to «make sure to not lose a single bit», but «make sure you can get your system back on quickly again in case of hardware/software failure».

homex, 11 months ago

@nixCraft

I don't rely on any FS, any storage server

Make many copy of data

Salix, 11 months ago

@nixCraft I have been using Btrfs since it's default in @opensuse (and also I have it in @turris_cz router and NAS), so I don't care about ZFS. 😉

klausman, 11 months ago

@nixCraft I've been using BTRFS for 15 years or thereabouts now. Haven't lost a single bit to it.

But this filesystem anecdata is just like people talking about harddrive models: for any given one, in a group there will be one person who has never had a spot of bother, and another who has only horror stories about said model.

At least we all agree that ReiserFS was a mistake.

known_as_bmf, 11 months ago

@nixCraft genuinely interested: why?

I don't know much about either FS but opensuse (which I use) comes out of the box with btrfs.