Re: [4.7.2] btrfs_run_delayed_refs:2963: errno=-17 Object already exists

[Marc Joliet <marcec@gmx.de>]

[-- Attachment #1: Type: text/plain, Size: 2981 bytes --]

On Montag, 6. März 2017 00:53:40 CET Marc Joliet wrote:
> On Mittwoch, 1. März 2017 19:14:07 CET you wrote:
> >  In any
> > 
> > case, I started btrfs-check on the device itself.
> 
> *Sigh*, I had to restart it, because I forgot to redirect to a file and
> quite frankly wasn't expecting this flood of output, but here's a summary
> of the output after about 2 days:
> 
[snip old output]

OK, it finished last night.  Here's the summary again:

% wc -l btrfs_check_output_20170303.log 
3028222 btrfs_check_output_20170303.log
% grep -v "backref lost" btrfs_check_output_20170303.log | grep -v "check \
(leaf\|node\) failed" | grep -v "lost its parent" | grep -v "referencer count" 
checking extents
ERROR: block group[3879328546816 1073741824] used 1072840704 but extent items 
used 1129164800
ERROR: block group[4163870130176 1073741824] used 1072259072 but extent items 
used 0
ERROR: block group[4223999672320 1073741824] used 1073664000 but extent items 
used 1074188288
ERROR: block group[4278760505344 1073741824] used 1073377280 but extent items 
used 1073901568
ERROR: block group[4406535782400 1073741824] used 1073627136 but extent items 
used 0
ERROR: extent [3830028140544 4096] referencer bytenr mismatch, wanted: 
3830028140544, have: 3826183847936
ERROR: errors found in extent allocation tree or chunk allocation
checking free space cache
checking fs roots
Checking filesystem on /dev/sdb2
UUID: f97b3cda-15e8-418b-bb9b-235391ef2a38
found 892572778496 bytes used err is -5
total csum bytes: 860790216
total tree bytes: 36906336256
total fs tree bytes: 34551476224
total extent tree bytes: 1230610432
btree space waste bytes: 7446885892
file data blocks allocated: 16359581663232
 referenced 2358137831424

> That's right, slowly approaching 1.5 million lines of btrfs-check output!
> That's *way* more than I ran it the last time this error happened a few
> weeks ago.

As can be seen above, that ballooned to over 3 million lines.  Since the 
output is 4.2 MB, even after XZ compression, I put it up on my Dropbox, just 
in case it's interesting to anybody:

https://www.dropbox.com/s/h6ftqpygfr4vsks/btrfs_check_output_20170303.log.xz?
dl=0

Since this is my backup drive, and the second time within a month that it had 
problems like this, *and* I've got both the btrfs-image dump and btrfs-check 
output, I'm going to go ahead and reformat, so that my three computers are 
finally backed up again.

Oh, and for what it's worth, I did test against a 4.8 kernel, and pretty much 
immediately got the "forced RO" error, just like with 4.9.  I didn't try 
anything older (or newer).

As a last step, I'll probably collect all information I have and post it to 
bugzilla when I have a chance, since others might hit it, too (Kai did before 
me, after all).

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

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 833 bytes --]