Hi,
On 2017-05-09 03:49, Adam Borowski wrote:
> Write hole is pretty nasty for metadata (likely to cause total filesystem
> loss) but when on -draid{5,6} -mraid{1,10} it's nowhere as bad. So for 4.12
> it might be ok to put up big warnings only for metadata. On the other hand,
> data loss limited to 1-2 files is still data loss -- CoW is supposed to
> never damage files already written.
>
> A real fix is obviously better than slapping on warnings.
The write hole is a real concern ? Only in the last year the linux MD raid implementation has gained a journal to avoid this problem. This means the in the last 15-years (or even more) this problem was here but its severity was "acceptable".
What I am trying to say, is that until the kernel 4.12, btrfs had several bugs which prevent to work even the basic raid5/6 functionalities (i.e. rebuild). This is a thing that the user should be warned. Because these kinds of bugs are unexpected by a "stable filesystem".
But the raid5/6 write hole is "defect" of all raid5/6 implementation. Until ZFS and the last iteration of MD, the real/only mitigation was a battery backup. In this BTRFS is not worse (nor better) than its competitor (xfs/ext3,4....). I am inclined to think that a warning for the write hole is a bit excessive.
BR
G.Baroncelli
--
gpg @keyserver.linux.it: Goffredo Baroncelli <kreijackATinwind.it>
Key fingerprint BBF5 1610 0B64 DAC6 5F7D 17B2 0EDA 9B37 8B82 E0B5
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
