[BUG] Please improve the following message "No space left on the device"

[elMor3no]

Hello,

I am helping to package rdiff-backup for Debian. Reviewing the reported bugs I found this request from a user.

Package: rdiff-backup
Version: 1.2.1-1
Severity: normal

Hi,

my system has 1 Gig of /tmp, and rdiff-backup happily fills that up
and then fails. To make things worse, it fails with a generic "No
space left on device" error and frees up /tmp again before dumping me
back to my shell, leaving absolutely no trace that it wasn't the
backup destination (which has a couple of hundred gig still free) that
was full.

It should _AT LEAST_ say that it was /tmp with no space left.
Preferably, it would not dump gigabytes of work data in /tmp - heck,
who has a /tmp this large?

Greetings
Marc

-- System Information:
Debian Release: lenny/sid
  APT prefers unstable
  APT policy: (500, 'unstable')
Architecture: i386 (i686)

Kernel: Linux 2.6.26.5-scyw00225 (PREEMPT)
Locale: LANG=de_DE.UTF-8, LC_CTYPE=de_DE.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/bash

Versions of packages rdiff-backup depends on:
ii  libc6                         2.7-13     GNU C Library: Shared libraries
ii  librsync1                     0.9.7-5    rsync remote-delta algorithm libra
ii  python                        2.5.2-2    An interactive high-level object-o
ii  python-support                0.8.6      automated rebuilding support for P

Versions of packages rdiff-backup recommends:
pn  python-pylibacl               <none>     (no description available)
pn  python-pyxattr                <none>     (no description available)

rdiff-backup suggests no packages.

-- no debconf information

As the request implies changes in the operation of the project, I decided to report the request here.

You can find more details in https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=500399

[ericzolf]

That's for a shit old version. Do you know if it's still the case?

[elMor3no]

Sadly I dont get any reply from the original of this threat https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=500399
Probably this error dont exist in recent version...

[ericzolf]

OK, I could reproduce the issue. It happens at least when you do a restore from a past increment, the diff is done in the temporary directory, and can fail if it's too small.

rdiff-backup --tempdir /mnt --remote-tempdir /mnt --api 201 --force restore --at 2B /tmp/bak /tmp/to
WARNING: Target path /tmp/to exists and isn't empty, content might be force
         overwritten by restore
WARNING: Repository is locked by file /tmp/bak/rdiff-backup-data/lock.yml,
         another action is probably on-going. Enforcing anyway at your own risk
NOTE:    Extended attributes not supported by filesystem at path
         /tmp/to/rdiff-backup.tmp.0 due to exception '[Errno 95] Operation not
         supported'
NOTE:    Starting restore operation from source path /tmp/bak to destination
         path /tmp/to
WARNING: Exception '[Errno 28] No space left on device' raised of class '<class 'OSError'>':
  File "/home/ericl/Comp/rdiff-backup/build/lib.linux-x86_64-3.10/rdiff_backup/robust.py", line 94, in check_common_error
    return function(*args)
  File "/home/ericl/Comp/rdiff-backup/build/lib.linux-x86_64-3.10/rdiffbackup/locations/_repo_shadow.py", line 1910, in get_fp
    current_fp = self._get_first_fp()
  File "/home/ericl/Comp/rdiff-backup/build/lib.linux-x86_64-3.10/rdiffbackup/locations/_repo_shadow.py", line 2025, in _get_first_fp
    rpath.copyfileobj(fp, current_fp)
  File "/home/ericl/Comp/rdiff-backup/build/lib.linux-x86_64-3.10/rdiff_backup/rpath.py", line 1546, in copyfileobj
    outputfp.write(inbuf)

ERROR:   Could not complete restore due to exception '[Errno 28] No space left
         on device'
ERROR:   Action restore failed on step run