|
 lock file for backup directory
Sebastien Maret <smaret < at > umich.edu>
wrote the following on Thu, 26 Jan 2006 11:26:43 -0500
Listing the increments in a backup directory while an other backup is in
progress give the following error:
rdiff-backup -l host::/path/to/backup
Fatal Error: Previous backup to /path/to/backup seems to have failed.
Rerun rdiff-backup with --check-destination-dir option to revert directory to state before unsuccessful session.
The message is confusing, since the previous backup hasn't failed,
but it just hasn't finished yet.
Beside, I am not sure if doing other operation (like delete old
increments) while a backup is in progress is safe. If not, it might be
useful to implement a lock system, to prevent two different
rdiff-backup processes to modify a backup directory simultaneously.
In general it's not a good idea to have two rdiff-backup processes
working on the same directory. I've changed the error message to:
The previous backup to %s is still in progress or has failed.
You can rerun rdiff-backup with --check-destination-dir option to revert
the directory to the state it was in before an unsuccessful session.
However,
*** -------------------------------------------------------------------- ***
*** If an rdiff-backup process is currently running, ***
*** DO NOT RUN --check-destination-dir OR OTHERWISE MODIFY THE DIRECTORY ***
*** -------------------------------------------------------------------- ***
which is hopefully clearer, and stresses the important parts :-)
Recent development versions of rdiff-backup will abort by default if
they think another rdiff-backup process is writing to the same
directory.
--
Ben Escoto
|
