[rdiff-backup-users] failure to regress, not enough space but there is

[Greg Troxel]

I have been doing rdiff-backup of a filesystem that's around 400 GB to
external disks for about 2 years.  As I got more bits, the free space on
the 1T backup disk (which has other stuff) got small, perhaps around
100G.  After a failed backup (due to me killing it, or something going
wrong, but nothing really all that odd), trying to do another one
(regressing) led to a complaint about no available disk space.

So I got a 2T disk, and rsync'd the rdiff-backup backups over.  Then,
backups of other filesystems happened fine, and I got the same
out-of-space error.  I wrote a script to run df once per second and log
it, and there was no sign of any increase in disk space usage at all -
450G is used, and there is over 1T free.

My system is NetBSD 5, i386, rdiff-backup 1.2.8, python 2.7, with UFS2
filesystems.  (UFS2 is basically the Berkeley Fast Fileystem, and very
stable.)

Here's the output from

"rdiff-backup --check-destination-dir /path/to/rdiff-target", which took
about 30 minutes to run, mostly cpu-bound:

Traceback (most recent call last):
  File "/usr/pkg/bin/rdiff-backup", line 30, in <module>
    rdiff_backup.Main.error_check_Main(sys.argv[1:])
  File "/usr/pkg/lib/python2.7/site-packages/rdiff_backup/Main.py", line 304, 
in error_check_Main
    try: Main(arglist)
  File "/usr/pkg/lib/python2.7/site-packages/rdiff_backup/Main.py", line 324, 
in Main
    take_action(rps)
  File "/usr/pkg/lib/python2.7/site-packages/rdiff_backup/Main.py", line 282, 
in take_action
    elif action == "check-destination-dir": CheckDest(rps[0])
  File "/usr/pkg/lib/python2.7/site-packages/rdiff_backup/Main.py", line 872, 
in CheckDest
    dest_rp.conn.regress.Regress(dest_rp)
  File "/usr/pkg/lib/python2.7/site-packages/rdiff_backup/regress.py", line 71, 
in Regress
    for rf in iterate_meta_rfs(mirror_rp, inc_rpath): ITR(rf.index, rf)
  File "/usr/pkg/lib/python2.7/site-packages/rdiff_backup/rorpiter.py", line 
281, in __call__
    last_branch.fast_process(*args)
  File "/usr/pkg/lib/python2.7/site-packages/rdiff_backup/regress.py", line 
268, in fast_process
    if rf.metadata_rorp.isreg(): self.restore_orig_regfile(rf)
  File "/usr/pkg/lib/python2.7/site-packages/rdiff_backup/regress.py", line 
290, in restore_orig_regfile
    tf.write_from_fileobj(rf.get_restore_fp())
  File "/usr/pkg/lib/python2.7/site-packages/rdiff_backup/restore.py", line 
488, in get_restore_fp
    return robust.check_common_error(error_handler, get_fp)
  File "/usr/pkg/lib/python2.7/site-packages/rdiff_backup/robust.py", line 32, 
in check_common_error
    try: return function(*args)
  File "/usr/pkg/lib/python2.7/site-packages/rdiff_backup/restore.py", line 
468, in get_fp
    Rdiff.write_patched_fp(current_fp, delta_fp, new_fp)
  File "/usr/pkg/lib/python2.7/site-packages/rdiff_backup/Rdiff.py", line 73, 
in write_patched_fp
    rpath.copyfileobj(librsync.PatchedFile(basis_fp, delta_fp), out_fp)
  File "/usr/pkg/lib/python2.7/site-packages/rdiff_backup/rpath.py", line 64, 
in copyfileobj
    outputfp.write(inbuf)
IOError: [Errno 28] No space left on device


Any clues?  Could it be using too much space in /tmp or /var/tmp?
But I saw no kernel messages about them being out of space.