I really like the system. Yet now the 2nd time I had to fix a full partition.
Today I came back after some weeks and did a big upgrade. Download was about 9 GB roughly.
While the ugprade was in progress the btrfs volume got full. The data part of btrfs volume was full.
I got it cleaned up by deleting old snapshots and data and with balancing I finally got it running again.
Is the default setting not aggresive enough for my setting? I have a 250 GB drive and 200 GB is actually used. So maybe there is not enough head room ?
Currently I would switch back to ext4 again. But maybe some people with more experience with btrfs can help me understand how I can keep a working system and maybe also get a notification that the drive is running full.
I make one occasionally if I haven’t been updating for a very long time and there is a long list of packages to be updated to avoid the risk of a potential breakage.
For my use case, a pre and post snaphot for every install, update, remove is overkill.
I don’t do much experimentation and extensive tweaking/tinkering either so KISS it is.
Currently i’m using just north of 70 gigs and for the average user including snapshots that’s a fairy big base with whatever apps you install on top of the OS install. The only real exceptions are video editing, audio editing, gaming. With even only 50 gigs free that 9 gig update should not have run you out of space even with creating snapshots every so often. That said I would go through and find any files you don’t actually need and move them to another drive so you can go through them, and make sure anything you don’t want to lose is on another drive.
Ok, I’m spoilt, but I don’t like running into that situation, so i buy more storage. My working daily driver partitions are at least 500 GB, and when they hit 70% I start cleaning house. I’ve never ran out of space in years.
It’s generally good for performance to leave about 20% free space (including whatever space is taken up by snapshots)
How much of that 200GB is used in areas that are actually snapshot vs. just stuff in /home, /var/cache, etc? It’s a smaller than ideal drive but with reasonable snapshot limits should be fine unless you have a ton of flatpaks, dockers, VMs etc that install into snapshot locations by default.
If you routinely go a long time between updates you should drop the number of retained snapshots (and maybe also expire them with time?)
Good idea switching to ext4, been using it since back in 2007 (I think it was ext3 back then). Anyway, use Timeshift to take one snapshot at startup, keep as many as you want (I keep 2).
Thank that was the biggest insight. I will try to adjust and have a look what files can be moved to the other drives or if I move a partition to a different drive.
Either way I will most likely switch to ext4 on the next installation just to have less complexity on the filesystem level.
Ext4 is perfectly suitable for “modern OS” because it’s fast and reliable. If I see the issues in this forum people have with BTRFS I gladly stick with Ext4 til end of time. Most people don’t need fancy stuff, just a file system that works.
In the end it’s a matter of taste and luckily we have that choice.
BTRFS is for sure on a level that is reliable in terms not corrupting data on its own.
BTRFS provides features and with it the complexity that not everyone wants. Some might think this complexity makes it more difficult to handle. Maybe that is what Oli meant.
Either way I think further discussions on this doesn’t make sense. Everyone made their statement from their perspective.