Learn how easy it is to sync an existing GitHub or Google Code repo to a SourceForge project! See Demo
As far as I can see, dar seems to finish its current backup even when it detects that an inode/file has changed during the process.
An option to make it stop immediately would be really useful.
next major release of dar (actually in pre-release phase, see http://dar.linux.free.fr/pre-release) can retry saving inode that changed (up to a maximum time and/or maximum amount of duplicated data). It also flags in the archive an inode that has changed, if it could not resave it without having it changing (a warning is also shown in that situation)
At restoration, such inode can be ignored (not restored), restored with confirmation (default), or restored without difference with other inode.
you also have to possibility to launch the command of your will before some given files or directory are to be saved and once the have been saved (like stopping a database and restarting it after backup).
Instead, having dar stopped abruptly as soon as a file has been changed during backup may lead one to think the backup has completed while it had not.
You have now all mean to avoid a file to change during backup, and if that cannot be avoided, you have to possibility to let dar retry the backup of such files. To my point of view, the feature you request is dangerous and not useful.
Don't you think so ?
The feature you describe for the next release has a much finer grain that what I proposed. (My intend was to restart the whole backup on such error, assuming that an abrupt stop would return an adequate error code.) So it's perfect to me! :-)
This Tracker item was closed automatically by the system. It was
previously set to a Pending status, and the original submitter
did not respond within 14 days (the time period specified by
the administrator of this Tracker).