Used it since 2011? I think, on various laptops. Used snapshotting frequently, and RAID1. No stability issues at all.
(I've had some very minor issues with ext3 in the past, massive corruption caused by LVM, lots of empty files with XFS, etc. btrfs did much better than any of these in the stability department.)
Performance has improved recently. Apparently, laptop SSDs can cause massive pauses when a sync is done? It used to be that the whole system froze for a long time; with kernel 4.8 and newer versions of btrfs, it seems that the pauses caused by SSDs does not have as severe an effect as they previously did.
Using it over a network (NBD) has been stable, but with a large number of files or blocks (not sure which) in the filesystem, it all of a sudden gets extremely slow. Something I would like to see improved (here's to hoping...).
Storing virtual machines on it works very poorly. Say that you install KVM+QEMU on a host system, format the harddrives with btrfs, disable COW with chattr, create a virtual hard disk (in RAW mode, no QCOW), then install an OS and enable LUKS and btrfs on top of that in the virtual machine. This, for some reason, causes massive fragmentation of the virtual machine hard drive file. It's something like a 4 TB disk would have 4 billion fragments, every single time. Multiply by a number of virtual machines, and you will quickly get tired of pausing virtual machines, copying the entire virtual harddrive from one file to another in the host system to get rid of the fragmentation, and un-pausing the virtual machines. A very big issue, so don't do this.. Not sure what the best filesystem (apart from VMFS) is for storing virtual harddrive images...
I'm curious about your NBD performance issues. When you say large number of blocks, are you talking a large filesystem? The free space cache tends to be a performance bottleneck for large filesystems, you might want to try out space_cache=v2.
I'm using it without data loss incidents for over 1.5 years now. On SSD drive I have encrypted volume. On top of it I have LVM volume. Ot top of it BTRFS partition. I use compress-force=lzo,autodefrag mount options. Works well for personal computer use case. I always follow latest kernel from Debian backports repo. I red that some RAID modes are implemented poorly it BTRFS still.
(I've had some very minor issues with ext3 in the past, massive corruption caused by LVM, lots of empty files with XFS, etc. btrfs did much better than any of these in the stability department.)
Performance has improved recently. Apparently, laptop SSDs can cause massive pauses when a sync is done? It used to be that the whole system froze for a long time; with kernel 4.8 and newer versions of btrfs, it seems that the pauses caused by SSDs does not have as severe an effect as they previously did.
Using it over a network (NBD) has been stable, but with a large number of files or blocks (not sure which) in the filesystem, it all of a sudden gets extremely slow. Something I would like to see improved (here's to hoping...).
Storing virtual machines on it works very poorly. Say that you install KVM+QEMU on a host system, format the harddrives with btrfs, disable COW with chattr, create a virtual hard disk (in RAW mode, no QCOW), then install an OS and enable LUKS and btrfs on top of that in the virtual machine. This, for some reason, causes massive fragmentation of the virtual machine hard drive file. It's something like a 4 TB disk would have 4 billion fragments, every single time. Multiply by a number of virtual machines, and you will quickly get tired of pausing virtual machines, copying the entire virtual harddrive from one file to another in the host system to get rid of the fragmentation, and un-pausing the virtual machines. A very big issue, so don't do this.. Not sure what the best filesystem (apart from VMFS) is for storing virtual harddrive images...