On Fri, 25 Sep 2015 09:12:15 -0400 Rich Freeman <r-btrfs@xxxxxxxxxxxxxxxxxx> wrote: > I'll just say that my btrfs stability has gone WAY up when I stopped > following this advice and instead followed a recent longterm. Right > now I'm following 3.18. There were some really bad corruption issues > in 3.17/18/19 that burned me, and today while considering moving up to > 4.1 I'm still seeing a lot of threads about issues during balance/etc. > I still run into the odd issue with 3.18, but not nearly to the degree > that I used to. > > Now, I would stick with a recent longterm. The older longterms go > back to a time when btrfs was far more experimental. Even 3.16 > probably has a lot of issues that are fixed in 3.18. Absolutely that! I was pondering whether or not to chime in with my praise of "longterm" as far as Btrfs stability goes, but apparently it's not just me who uses it. In my experience 3.18 just works* and is very stable, and before that it was 3.14, which by luck(?) happened to go longterm IIRC just before Btrfs transitioned to "kernel worker threads" in 3.15 (and that caused ALL sorts of trouble initially). [*] at least in a relatively simple scenario -- with snapshots, but without using any of the multi-device features or stuff such as qgroups or send/receive. -- With respect, Roman
Attachment:
signature.asc
Description: PGP signature
