Re: [rdiff-backup-users] restoring files

[Andrew K. Bressen]

Mal Beaton <address@hidden> writes:
> "/home/tretkowski/rdiff-backup/rdiff-backup-0.13.4/debian/rdiff-backup/usr/lib/python2.2/site-packages/rdiff_backup/connection.py",
> line 230, in _get
> rdiff_backup.connection.ConnectionReadError: Truncated header
> string (problem probably originated remotely)

this error in my experience usually means the network connection
has dropped unexpectedly, which isn't terribly useful since anything
causing the server process to suddenly die would close down the 
network connection. 

can the remote backup client successfully list increments on the server? 
also, try a network restore over ssh from the backup server to itself. 

could there be a permissions problem, such as the local user running the
restore not having permission to write to the restore destination?

are the remote backup and restore attempts both using the
same userid on the server? is this also the userid that can successfully
do a local restore on the server? do the permissions and ownerships
on the backups and restore target area look sane? 

if you can, see how far the ssh connection gets; 
does a login successfully happen on the server during the 
restore attempt? 

if none of this helps, try strace'ing the client restore to see what
is going on when things die, and if that still doesn't help, strace a
local restore on the backup server and see where the two differ. 

there should be some way to strace the server process invoked by the
remote client while it runs; I've never had to do that, but if I
needed to, I'd try some kludge like attaching strace to the sshd and
having it follow process forks (-f and -ff options), or replacing the
rdiff-backup executable with a bash script that runs the program under
strace (while passing all the arguments).


  --best of luck
  --akb