>>>>> On Tue, 28 Jun 2005 09:39:51 +0530, Siju George <[EMAIL PROTECTED]> said: Siju> DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws;
Siju> On 6/28/05, Martin Simmons <[EMAIL PROTECTED]> wrote: >> >>>>> On Mon, 20 Jun 2005 10:20:49 +0200, Arno Lehmann <[EMAIL PROTECTED]> said: >> Arno> Hello, >> Arno> Siju George wrote: >> >> >> On 6/20/05, Siju George <[EMAIL PROTECTED]> wrote: >> >> >> >>> Hi, >> >>> >> >>> >> >>> I am looking at Bacula >> >>> but the line >> >>> # Files deleted after a Full save will be included in a restoration. >> >>> in >> >>> http://www.bacula.org/rel-manual/Current_State_Bacula.html >> >>> under >> >>> Current Implementation Restrictions >> >>> makes me a little apprhensive. >> >>> Is Bacula actually not able to backup Directory structures?? >> >>> >> >>> i.e >> >>> >> >>> If in a Directory "dir1" I have "file1", "file2" & "file3" and I take a >> >>> full backup "backup1" on "Day1" the I delete "file2" from "dir1" and >> >>> take a differential backup "backup2". >> >>> >> >>> Now I should be able to restore "dir1" from "backup2" with only "file1" >> >>> and "file2" >> >> >> >> >> >> Sorry :-( this should read as "file1" & "file3" >> >> >> Arno> You _can_ restore file1 file2 and file3 in dir1. However, when you use Arno> bacula to manage a restore and tell it to make a complete restore it Arno> will restore all three files. Arno> If you tell bacula to only use the differential backup then, of course, Arno> file2 will not be restored. Usually you will not do this because it's Arno> baculas job to keep track which jobs and volumes it needs to restore :-) >> Arno> So, of course bacula saves and retrieves directory structure, but it Arno> does not keep track of deletions. >> Arno> The latter would require a complete compare of all directory entries to Arno> be backed up with what bacula has in its catalog and thus would be very Arno> resource intensive. >> >> Done correctly, it should be possible to do all the work in restore for >> filesystems that work properly. Backup just has to record the inode number >> for each changed file. >> Siju> Could you please explain how??? For each file saved in the backup, you record its full path and some unique id that remains unchanged during renaming (e.g. the inode number on UNIX). For each dir saved, you also record the names of the all files/dirs in that dir. You do both of these things for the initial full backup and subsequent incr/diff backups. During the restore of the full backup, you create a table mapping the unique ids to the full paths. Then during the restores of the incr/diff backups, you can search the table for each unique id being restored and decide if the file is new or renamed (or newly hard-linked). For directories you can check the list of names in that directory to detect deleted or renamed files. This only works on filesystems where renaming a file will change its ctime (or some other property) and where deleting a file will change the mtime (or some other property) of its directory. This is how ufsdump and star work. __Martin ------------------------------------------------------- SF.Net email is sponsored by: Discover Easy Linux Migration Strategies from IBM. Find simple to follow Roadmaps, straightforward articles, informative Webcasts and more! Get everything you need to get up to speed, fast. http://ads.osdn.com/?ad_id=7477&alloc_id=16492&op=click _______________________________________________ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
