[rdiff-backup-users] global name 'e' is not defined

[Michael Grant]

I get the following exception:

Found interrupted initial backup. Removing...
Traceback (most recent call last):
  File "/usr/local/bin/rdiff-backup", line 30, in ?
    rdiff_backup.Main.error_check_Main(sys.argv[1:])
  File "/usr/local/lib/python2.4/site-packages/rdiff_backup/Main.py",
line 309, in error_check_Main
    errmsg = robust.is_routine_fatal(exc)
  File "/usr/local/lib/python2.4/site-packages/rdiff_backup/robust.py",
line 77, in is_routine_fatal
    elif isinstance(exc, EnvironmentError) and e.errno == errno.ENOTCONN:
NameError: global name 'e' is not defined


On line 77 of robust.py, should this perhaps be exc.errno instead of e.errno?


I believe I managed to generate the above error because I am trying to
backup to a directory that has a mount point in it that is mounted
read-only.  I essentially have 4 partitions:

/
/backup
/home
/backup/home

/  is backed up to /backup/
/home is backed up to /backup/home/

after the backups, I do a 'mount -u -r /backup' and mount -u -r
/backup/home'.  This makes it very easy for someone to find a backup
just by pre-pending /backup to the path of the desired file.

The problem occurred because I had an interrupted backup in / and the
second time I went to do the backup, it tried to do an rm -r in
/backup, but when it came to my mount point of /backup/home, it tried
to remove it's contents too, but it being read-only, it wasn't allowed
and hence threw the error.

That leads to a second bug, when --exclude-other-filesystems is given,
rdiff-backup should also be careful to ignore other file systems in
the destination dir as well.  The justification for this is that if
you are not traversing a mount point in the source dir, you certainly
don't want to traverse a mount point in the destination dir.

A workaround for this is simply to umount /backup/home before doing
the backup to /backup.

Michael Grant


Michael Grant