[rdiff-backup-users] failing to fix backup

[Hamish Downer]

Hello

I've spend some time trying to fix one of our rdiff-backup backups and
after working through a number of problems I'm now getting the message:

    assert newrp.getinctype() == oldrp.getinctype() == 'snapshot'

To give some background, we're backing up about 300G of shared home
space from one NFS mount to another. The server running the backups is
running Debian 6 and rdiff-backup 1.2.8 The server is now 64 bit, as it
ran out of memory when trying to regress destination after an earlier
failure (it hit 3G of RAM and gave an OOM Error).

More than once, since the last successful backup I've followed the
advice at http://www.nongnu.org/rdiff-backup/FAQ.html#regress_failure
deleting a current_mirror file and moving a mirror_metadata file out of
the way.

The last few times it failed I was getting errors of the type:

    IOError: [Errno 23] Too many open files in system: 

But the file it named was different, and further along, each time, so I
re-ran it again, though often having to do the current_mirror/
mirror_metadata thing to get it to run.

The third last run gave me the message at the top of the file. I tried
running again with --verbosity 9 and got:

    Wed May 16 08:44:16 2012  Previous backup seems to have failed, regressing 
destination now.
    Wed May 16 08:44:17 2012  Warning, could not find mirror_metadata file.  
Metadata will be read from filesystem instead.
    Wed May 16 08:44:17 2012  Fatal Error: No metadata for time Mon May 14 
08:24:04 2012 (133 cannot regress

deleting current mirror and moving mirror metadata has got me back to
the assert error, which appeared twice, the second one being the end of
the file. The last few lines (of the 1.9G log file, verbosity 9) are:

Wed May 16 21:54:45 2012  Copying attributes from () to 
/mnt/backup/fen-apps/home
Wed May 16 21:54:45 2012  Setting time of /mnt/backup/fen-apps/home to 
1305115103
Wed May 16 21:54:45 2012  Touching 
/mnt/backup/fen-apps/home/rdiff-backup-data/extended_attributes.2012-05-16T10:11:27+01:00.snapshot
Wed May 16 21:54:45 2012  Touching 
/mnt/backup/fen-apps/home/rdiff-backup-data/access_control_lists.2012-05-16T10:11:27+01:00.snapshot
Traceback (most recent call last):
  File "/usr/bin/rdiff-backup", line 30, in <module>
    rdiff_backup.Main.error_check_Main(sys.argv[1:])
  File "/usr/lib/pymodules/python2.6/rdiff_backup/Main.py", line 304, in 
error_check_Main
    try: Main(arglist)
  File "/usr/lib/pymodules/python2.6/rdiff_backup/Main.py", line 324, in Main
    take_action(rps)
  File "/usr/lib/pymodules/python2.6/rdiff_backup/Main.py", line 280, in 
take_action
    elif action == "backup": Backup(rps[0], rps[1])
  File "/usr/lib/pymodules/python2.6/rdiff_backup/Main.py", line 343, in Backup
    backup.Mirror_and_increment(rpin, rpout, incdir)
  File "/usr/lib/pymodules/python2.6/rdiff_backup/backup.py", line 51, in 
Mirror_and_increment
    DestS.patch_and_increment(dest_rpath, source_diffiter, inc_rpath)
  File "/usr/lib/pymodules/python2.6/rdiff_backup/backup.py", line 245, in 
patch_and_increment
    cls.CCPP.close()
  File "/usr/lib/pymodules/python2.6/rdiff_backup/backup.py", line 491, in close
    metadata.ManagerObj.ConvertMetaToDiff()
  File "/usr/lib/pymodules/python2.6/rdiff_backup/metadata.py", line 640, in 
ConvertMetaToDiff
    newrp, oldrp = self.check_needs_diff()
  File "/usr/lib/pymodules/python2.6/rdiff_backup/metadata.py", line 629, in 
check_needs_diff
    assert newrp.getinctype() == oldrp.getinctype() == 'snapshot'
AssertionError


Where next?

Hamish