Re: btrfs send kernel error btrfs_compare_tree

[Duncan <1i5t5.duncan@cox.net>]

Travis Cross posted on Sat, 08 Mar 2014 20:35:16 +0000 as excerpted:

> The filesystem here was likely created with Linux 3.2 and hasn't seen
> much use for awhile, until today I mounted it to try to btrfs send off
> those volumes.
> 
> xaba noted this could be the result of some 3.2-era kernel bug.  He
> recognized the messages I was seeing.  If this is the case, and this
> sort of thing is common, it seems we might want to have a way of
> detecting this and trying to salvage the situation (particularly as
> Debian wheezy -- the last Debian stable release -- is on a 3.2 kernel).

Well, until 3.13 (IIRC) btrfs was officially experimental, with a very 
strongly worded warning on the kernel option activating it.  And even 
after that semi-stabilization (the wording still doesn't suggest fully 
stable) current wiki and mkfs.btrfs strongly encourage keeping current on 
your kernel if you're running btrfs, something kernel 3.2 definitely is 
*NOT*.

So I'd consider backing up the data and doing a clean mkfs.btrfs on the 
filesystem, starting over with a filesystem created with a post-eat-your-
babies-warning kernel.  I did that here recently, taking advantage of 
several of the newer btrfs disk-format features, and plan to do it again 
at least once more after a few more kernel cycles of code settling, just 
to be sure I'm not relying on something written by potentially still 
buggy and not yet entirely stable btrfs code.

Tho I expect the devs will try to salvage this specific situation and 
have a bug-fix for it.  But I know at least personally, I rest better 
knowing that none of my btrfs has been touched by that officially still 
very experimental code; they've all been redone with a newer kernel 
beyond that warning and haven't run anything older, and as I said, I plan 
to redo them again at least once more, as btrfs settles down further.

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman