Conversation

Notices

1. Embed this notice
   Ryan Castellucci :nonbinary_flag: (ryanc@infosec.exchange)'s status on Friday, 31-Jan-2025 18:02:07 JST Ryan Castellucci :nonbinary_flag:

   Is btrfs still terrible?

   • Embed this notice
     Adam Chovanec (staticnoisexyz@infosec.exchange)'s status on Friday, 31-Jan-2025 18:20:36 JST Adam Chovanec

     in reply to

     @ryanc default on fedora and arch afaik. Solid ob my laptop for a few years.

   • Embed this notice
     tsk (tasket@infosec.exchange)'s status on Friday, 31-Jan-2025 18:23:38 JST tsk

     in reply to

     @ryanc I'd say not. Reliability is far better than Ext4-on-Thin LVM, and its worst-case throughput scenarios (esp. large container disk image files) have simple workarounds.

     Its still fashionable to do benchmarks comparing Btrfs with plain XFS and Ext4. IMO those comparisons only tell you that copy-on-write has a cost. If you want copy-on-write features then Btrfs performance on Linux >6.1 kernels is solid.

     There are no more "out of metadata space" oopses as Btrfs will now automatically allocate more mdata space as needed.

     Of course btrfs-send is brilliant (better than zfs-send) although there are other backup tools like Wyng that can achieve a similar level of efficiency using the metadata Btrfs makes available.