Karl Cunningham a écrit : > Baptiste Malguy wrote: > >> Hello, >> My sequence: >> >> Step 1 >> - Ignore FileSet Changes = no >> - File = /home/test/d1 >> - Requesting an incremental backup. It is actually incremental (the full >> done previously with the same setting) >> >> Step 2 >> - Ignore FileSet Changes = yes >> - File = /home/test/d1 >> - Requesting an incremental backup. Still observe an incremental. >> - No new FileSet checksum >> >> Step 3 >> - Ignore FileSet Changes = yes >> - File = /home/test/d2 (d1 and d2 have diffrent contents) >> - Requesting an incremental backup. Still observe an incremental. (d2 is >> backuped) >> - No new FileSet checksum >> >> Step 4 >> - Ignore FileSet Changes = no >> - File = /home/test/d2 >> - Requesting an incremental backup. Still observe an incremental. (d2 is >> backuped _again_) >> - New FileSet checksum >> >> I'm a little surprised by the results. I'm not sure this is the designed >> behavior by the developers. > > > Baptiste -- > > Sorry, I forgot to copy the list in my first reply to this question. > Here it is... > > I won't address the question of whether a changing IgnoreFilesetChanges > should itself be deemed a fileset change. > > I don't know what the intent of the developers was, but I believe what > you are describing comes down to the question of which fileset is used > as a reference to detect changes. Two possible methods come to mind: > > 1. The most recent fileset used while IgnoreFilesetChanges was yes. > 2. The most recent fileset used regardless of the setting of > IgnoreFilesetChanges. > > Either method has its advantages and disadvantages. I think you were > expecting method 1 but what is implemented is method 2. I believe what > is implemented will satisfy most users' needs, but perhaps the method of > checking for changes needs to be more thoroughly described in the manual.
Yes, I expected proposal 1, with this detail: the most recent fileset _before_ IgnoreFileSetChanges was set. My guess is it does not pay attention if the FileSet read in the conf file has a matching entry in the database. However it indicates the most recent FileSet in the report. I don't think proposal 2 is what happened. If it was, it would not have backed up changes in d2/ made after IgnoreFilesetChanges was set to yes as there were not in the Files attribute of the last FileSet recorded in the database. > >> I'm wrong here. Only changes done in d2/ are backuped. Changes since >> when ? (last Full ? Last Inc ?) > > > If d2 is a file, with an incremental it would be backed up if the file > changed since last time it was backed up. If d2 is a directory, files > in it would be backed up if they changed since last time they were > backed up. Each file is examined individually. d2 was added to the FileSet after the Full dump. As a consequence, Bacula had no reference to this directory (in my case it's a directory, but nevermind) about its last backup. The doc says a file is backed up if its change (attributes) or modification (content) times are changed since the last Full/Diff/Incr job. Based on this, the files/dir in d2/, older than the last backup are _not_ saved. To do so, it would be necessary to also check if each file/dir has been previously backed up. Instead of this, by default, next backup is upgraded to a Full backup. PS: so you say "backed up" and not "backuped" ? I didn't know how to write it :) -- Baptiste MALGUY - IT Engineer SOLSOFT PGP Fingerprint: 2A90 B6D1 855F 1584 FE68 94AC CCD6 36D0 3DC7 DCB8 www.solsoft.com - phone: +33 1 47 15 04 00 - fax: +33 1 47 15 55 09 How secure networks are managed. ------------------------------------------------------- This SF.Net email is sponsored by xPML, a groundbreaking scripting language that extends applications into web and mobile media. Attend the live webcast and join the prime developer group breaking into this new coding territory! http://sel.as-us.falkag.net/sel?cmd=lnk&kid0944&bid$1720&dat1642 _______________________________________________ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users