Re: [gentoo-amd64] btrfs Was: Soliciting new RAID ideas

[Marc Joliet]

(Dammit, it seems that I've developed a habit of writing somewhat long-winded
emails :-/ . Sorry!)

Am Wed, 28 May 2014 08:29:07 +0100
schrieb thegeezer <thegeezer@×××××××××.net>:

> top man, thanks for detail and the tips !

I second this :) . In fact, I think I'll link to it in my btrfs thread on
gentoo-user.

I do have a question for Duncan (or anybody else who knows, but I know that
Duncan is fairly active on the BTRFS ML), though:

How does btrfs handle checksum errors on a single drive (or when self-healing
fails)?

That is, does it return a hard error, rendering the file unreadable, or is it
possible to read from a corrupted file?  Sadly, I don't remember finding the
answer to this from my own research into BTRFS before I made the switch (my
thread is here: [0]), and searching online now hasn't revealed anything; all I
can find are mentions of its self-healing capability.

I *think* BTRFS treats this as a hard error? But I'm just not sure.

(I feel kind of stupid, because I'm sure I saw the answer in some of the emails
on linux-btrfs that I read through via GMANE.)

I ask because I'm considering converting the 2TB data partition on my 3TB
external hard drive from NTFS to BTRFS [1] .  It primarily contains media
files, where random corruption is decidedly *not* the end of the world.
However, it also contains ISOs and other large files where corruption matters
more, but which are not important enough to land on my BTRFS RAID (on the other
hand, my music collection is ;-) ).

In any case, reconstructing a corrupted file can be fairly difficult: It might
involve re-ripping a (game) disk, or it might be something I got from a friend,
delaying file recovery until I can get it again, or the file might be a youtube
download (or a conference video, or something from archive.org, or ...) and I
have to track it down online again.  However, I might want to *know* that a file
is corrupt, so that I *can* reconstruct it if I want to.

The obvious answer, retrieving from backup, is difficult to implement, since I
would need an additional external drive for that.  Also, the files are not
*that* important, e.g., in the case of a youtube download, where most of the
time I delete the file afterwards anyway.

(It seems to me that the optimal solution would be to use some sort of NAS, with
a multi-device ZFS or BTRFS file system, in place of an external hard drive; I
expect to go that route in the future, when I can afford it.)

[0] http://thread.gmane.org/gmane.linux.gentoo.user/274236

[1] I used NTFS under the assumption that I might want to keep the drive Windows
compatible (for family), but have decided that I don't really care, since the drive is
pretty much permanently attached to my desktop (it also has an EXT4 partition
for automatic local backups, so removing it would be less than optimal ;-) ).

-- 
Marc Joliet
--
"People who think they know everything really annoy those of us who know we
don't" - Bjarne Stroustrup