AttributeError: 'Volume' object has no attribute 'live' (error with multiple filesystems)

[nefelim4ag]

Traceback (most recent call last):
  File "/usr/bin/bedup", line 9, in <module>
    load_entry_point('bedup==0.9.0', 'console_scripts', 'bedup')()
  File "/usr/lib/python3.3/site-packages/bedup/__main__.py", line 482, in script_main
    sys.exit(main(sys.argv))
  File "/usr/lib/python3.3/site-packages/bedup/__main__.py", line 471, in main
    return args.action(args)
  File "/usr/lib/python3.3/site-packages/bedup/__main__.py", line 197, in vol_cmd
    dedup_tracked(sess, volset, tt, defrag=args.defrag)
  File "/usr/lib/python3.3/site-packages/bedup/tracking.py", line 393, in dedup_tracked
    dedup_tracked1(ds, ofile_reserved, query)
  File "/usr/lib/python3.3/site-packages/bedup/tracking.py", line 459, in dedup_tracked1
    with open_by_inode(inode, ds.sess, query) as rfile:
  File "/usr/lib/python3.3/contextlib.py", line 48, in __enter__
    return next(self.gen)
  File "/usr/lib/python3.3/site-packages/bedup/tracking.py", line 403, in open_by_inode
    pathb = inode.vol.live.lookup_one_path(inode)
AttributeError: 'Volume' object has no attribute 'live'

[phiresky]

Same for me. When rerunning it gets a little further, then stops with the same error

[neopeak]

same problem here

[neopeak]

found a workaround: Delete the sqlite databases under share/bedup and restarted the process..

I might had an old sqlite db schema from a previous btrfs experiment lying around.

[phiresky]

thanks, this seems to have solved the problem for me also

[g2p]

I can reproduce this with Python 3 only (though @adamryczkowski / #41 saw this in Python 2). I'm guessing it has to do with the way strings are stored in the db.

[keturn]

got this with pypy 2.2.1. Not always! It did one FS fine, and got this partway through a run of another FS.

The one it gave this message on was specified by UUID.

and I'd interrupted a previous run with Ctrl-C during the scan phase, when it told me "/ is not the root, use a UUID instead." Don't know if interrupting it corrupted something.

kernel 3.13.0-16-generic #36-Ubuntu x86_64

[alanfairless]

I made a temporary workaround. https://github.com/alanfairless/bedup/commit/f8e56ed0d1827bb5c237a3db55f8c5ffbe312b8a

[phiresky]

I thought deleting the cache db and using the newest git version while not interrupting scanning fixes it, but still the same:

[tonal]

Also always reproduced:

$ sudo rm ~/.local/share/bedup/db.sqlite*
$ sudo ~/.local/bin/bedup dedup --defrag /dev/sda3 
Scanning volume {e0fcba5f-71a4-404e-9a25-a11ccc745f01} generations from 0 to 69077, with size cutoff 8388608
...
01:04.0 Size group 105/105 sampled 348 hashed 65 freed 95994243
00.00 Committing tracking stateNo handlers could be found for logger "sqlalchemy.pool.SingletonThreadPool"
00.15 Committing tracking state

$ sudo ~/.local/bin/bedup dedup --defrag /dev/sda4
Scanning volume {7f788e51-7f6b-470c-90a9-59fcb980c05c} generations from 0 to 7, with size cutoff 8388608
...
Deduplicating filesystem {7f788e51-7f6b-470c-90a9-59fcb980c05c}
01:00.6 Size group 23/48 sampled 48 hashed 10 freed 0
Traceback (most recent call last):
  File "/home/nastja/.local/bin/bedup", line 9, in <module>
    load_entry_point('bedup==0.9.0', 'console_scripts', 'bedup')()
  File "/home/nastja/.local/lib/python2.7/site-packages/bedup/__main__.py", line 483, in script_main
    sys.exit(main(sys.argv))
  File "/home/nastja/.local/lib/python2.7/site-packages/bedup/__main__.py", line 472, in main
    return args.action(args)
  File "/home/nastja/.local/lib/python2.7/site-packages/bedup/__main__.py", line 198, in vol_cmd
    dedup_tracked(sess, volset, tt, defrag=args.defrag)
  File "/home/nastja/.local/lib/python2.7/site-packages/bedup/tracking.py", line 394, in dedup_tracked
    dedup_tracked1(ds, ofile_reserved, query)
  File "/home/nastja/.local/lib/python2.7/site-packages/bedup/tracking.py", line 460, in dedup_tracked1
    with open_by_inode(inode, ds.sess, query) as rfile:
  File "/usr/lib/python2.7/contextlib.py", line 17, in __enter__
    return self.gen.next()
  File "/home/nastja/.local/lib/python2.7/site-packages/bedup/tracking.py", line 404, in open_by_inode
    pathb = inode.vol.live.lookup_one_path(inode)
AttributeError: 'Volume' object has no attribute 'live'

Versions: OS - Kubuntu 14.04 with all update

 13:31:58  
$ pip install -U --user bedup
Requirement already up-to-date: bedup in ./.local/lib/python2.7/site-packages
Requirement already up-to-date: alembic in ./.local/lib/python2.7/site-packages (from bedup)
Requirement already up-to-date: cffi>=0.4.2 in ./.local/lib/python2.7/site-packages (from bedup)
Requirement already up-to-date: pyxdg in /usr/lib/python2.7/dist-packages (from bedup)
Requirement already up-to-date: SQLAlchemy in ./.local/lib/python2.7/site-packages (from bedup)
Requirement already up-to-date: contextlib2 in ./.local/lib/python2.7/site-packages (from bedup)
Requirement already up-to-date: Mako in ./.local/lib/python2.7/site-packages (from alembic->bedup)
Requirement already up-to-date: pycparser in ./.local/lib/python2.7/site-packages (from cffi>=0.4.2->bedup)
Requirement already up-to-date: MarkupSafe>=0.9.2 in ./.local/lib/python2.7/site-packages (from Mako->alembic->bedup)
Cleaning up...

[lilydjwg]

I've finally got some ideas on this issue.

When I only dedup one filesystem it works perfectly, but as I use the same database for multiple filesystems this issue may occur, because there are two filesystems with the same volumn id 5 (the root subvolume). In the example below, I'm deduping fs_id=2, but there is one Inode with fs_id=1 got creeped in.

>>> inodes
[Inode(ino=10930, volume=109), Inode(ino=48100, volume=5), Inode(ino=96584, volume=109)]
>>> inodes[1].vol.fs_id
1
>>> inodes[2].vol.fs_id
2

This explains why in #46 someone says removing the database works around this (but it doesn't work for me). The "real" workaround might be use one database per filesystem. I'll try that later.