3 ms·
Using BTRFS was a mistake. The interval trees became corrupted and the repair tools were useless and made the situation even worse. I was tempted to use it bec
by ok123456 4y ago
Using BTRFS was a mistake. The interval trees became corrupted and the repair tools were useless and made the situation even worse.
I was tempted to use it because it had copy on write de-duplication. But, after that experience, I'm sticking to EXT4.
- GauntletWizard 4y agoI have heard too many stories of data corruption on BRTFS, and it saddens me, because initially it looked really good. Meanwhile my 10 year old ZFS volume has had it's RAID-6 Drives replaced fully 3 times, and never lost data.
- sph 4y agoI keep hearing this, with no reference whatsoever to _when_ you had issues with it. Meanwhile distros like Fedora or Synology on their NAS have been running BTRFS for a while, and there is not much evidence of these legendary data corruption moments. Last time I posted this comment on a similar anecdote, I got no response. Do you have more evidence to show how fragile btrfs is in 2022? Because from where I stand it they read like horror stories from users when the btrfs code was very new and unstable, and then echoed ad nauseam by other people. See the sibling comments for example.
- ok123456 4y agoThis was last year. I'd be surprised if it got much better in 12+ months.
- sph 4y agoWere you using BTRFS in RAID mode by any chance? Have you opened a kernel bug report?
- ok123456 4y agoNo. Report what? The stack trace saying the extent tree is bad? There are always going to be consistency errors when you're dealing with filesystems, especially when you're non-ECC memory. They need to be robust enough that the fsck tools can resolve that. From my experience with BTRFS, doesn't do that.