Hi, Krsnendu dasa wrote on 21.01.2007 at 13:45:55 [Re: [BackupPC-users] Avoiding long backup times]: > On 20/01/07, Holger Parplies <[EMAIL PROTECTED]> wrote: > > > I read, that backuppc stores every identical file only one time. > > > > Basically true, but BackupPC needs to determine that the file is identical > > to > > something and to what. Sparing you the transfer (if possible with reasonable > > cost) is rsyncs job. Doing so comes at a cost in terms of CPU usage, so > > you've got the option of using tar if bandwidth is cheaper than CPU power. > > In your case, it obviously is not. > > > > > [...] What happens with an unchanged file in a full backup? > > > > tar and smb will transfer it, rsync will not (presuming you mean > > "unchanged" as in "same name, same content"). > > > > For a changed file, rsync will try to speed up the transfer. If you append a > > few bytes to a large file, tar/smb will transfer the whole file (even on an > > incremental), while rsync will (basically) transfer some checksums and the > > few bytes only (on full and incremental backups). > > > So... > > With tar and smb the file is transfered then it is checked to see if > it is the same as a file already in the pool. If it is already in the > pool a link is created and the copied file is deleted. > > Whereas rsync checks if the file is the same as another in the pool > before transferring?
No. Actually, I was thinking of *native rsync* in what I wrote. As I read the code (or the comments, rather ;-), the same basically applies to BackupPC/File::RsyncP, though there are two additional challenges worth mentioning: compression of the local file and pooling. As Les pointed out, the point of reference is always the same file in the previous full backup, not any file in the pool that may (later) turn out to have identical contents. This reference file may be compressed. In any case, it's part of a previous backup, so the contents cannot be changed "in place". As I understand it, BackupPC will create a local directory tree containing the required files (all for a full backup, only the changed ones for an incremental), much as native rsync would with the '--copy-dest' or '--compare-dest' options respectively (meaning, an identical file will be copied locally, not transfered, and a changed file will be partially transfered if possible), except that BackupPC will already compress the local file if compression is turned on. Later on, BackupPC_link will replace files in this directory tree by hardlinks to pools files or add them to the pool as appropriate. I like the idea of taking *any* identical file from the pool as reference though. I don't know if it is possible (i.e. whether the remote rsync will transfer one checksum covering the whole file before the local part needs to commit to selecting one file), but it certainly would be a worthwhile speedup. Regards, Holger ------------------------------------------------------------------------- Take Surveys. Earn Cash. Influence the Future of IT Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys - and earn cash http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV _______________________________________________ BackupPC-users mailing list BackupPC-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/backuppc-users http://backuppc.sourceforge.net/
