On Sun, Oct 5, 2025, at 6:08 AM, Javier Perez wrote: > Is there any link, tutorial, etc that explains what kind of btrfs maintenance > we are supposed to be doing?
I do no regular maintenance other than backup of important data. Perhaps a few times a year I will scrub the file system, and check dmesg for errors. Upstream maintainer also maintains 'btrfsmaintenance' which is packaged in Fedora. The 'btrfs-balance.timer' can optionally be enabled. This periodically runs a filtered balance. On the one hand prematurely running out of space is a bug, and in the unlikely case you experience it, don't try to fix it, we need a detailed bug report instead. On the other hand bug reporting is voluntary and it's reasonable to just avoid issues. And hence btrfs-balance.timer from the btrfsmaintenance package. For those on the fence, I don't have a hard and fast rule of thumb but if you push storage beyond 90% full you may want to use it. (I do push some less active storage well beyond 90%, yet still do not balance at all.) There is a newer feature built-in to the kernel called dynamic and periodic reclaim. They are safe to enable via the sysfs interface on a per block group basis. There's a proposal to enable them for data block groups by default in a future kernel, but is not yet merged. Note from the description it doesn't become active until unallocated space goes below 10G. This is the last reported type of space when using 'btrfs filesystem usage /' https://lore.kernel.org/linux-btrfs/52b863849f0dd63b3d25a29c8a830a09c748d86b.1752605888.git.bo...@bur.io/ It's hoped this will obviate the btrfs-balance.timer but there's no conflict or incompatibility using both. Fedora Btrfs matrix room to discuss specifics https://matrix.to/#/#btrfs:fedoraproject.org > Long time btrfs user and have never knowingly performed any maintenance on my > btrfs filesystem except balancing after converting from single HDD to raid 1 > my system. > I wonder what I have been missing (ignorance is bliss :( ) > Yep. This is fine. I don't have a way to track this with any real basis except feels ... but I'm seeing fewer space related issues with Btrfs than memory bit flips and drive firmware bugs. So, not knowing what problem you might run into, backup your important data so you don't have to worry. Any backup will do. Whatever you're familiar with and will use. I myself prefer btrfs snapshot replication using send+receive. The incremental send is ridiculously cheap, almost magic. No deep traversal required on either source or destination. If you have many files with few changes, you will see a performance improvement just by entirely avoiding the scan needed to determine what few files need backing up. If you're not noticing this scan for file changes, then don't worry about it. Just keep doing what you're doing. Chris Murphy
-- _______________________________________________ users mailing list -- [email protected] To unsubscribe send an email to [email protected] Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/[email protected] Do not reply to spam, report it: https://pagure.io/fedora-infrastructure/new_issue
