> Mike McMullen wrote: > > Hi all, > > > > I have a detail question on rsync. If you don't use --whole-file > > how does rsync copy a file especially a binary or a large mail > > box? > > I'm not sure the full purpose of --whole-file. Maybe if you copy of lots of > teeny-tiny files, it may eliminate the wasted time of diff checks. But for > large files, I don't see the point of using --whole-file. And if the > destination file doesn't exist, well rsync does a "whole-file" copy anyway. > > rync just syncs bits by the divide and conqueror method regardless of the > file being text of binary - it's really fast. > > Don't use --whole-file if you're doing syncs over slow wan links like a T1. > -eric wood > Hi Eric, In our scenario we do have lots and lots of scanned tiffs and pdfs so --wholefile on a private server network cut down on load. The main reason for --wholefile was my concern over data corrupted if incremental changes didn't work or more likely didn't work like I thought it did. I did some sample rsyncs to the off site server and pulled the data over and verified that it was ok. Without --wholefile it flew over the T1 and the load wasn't bad enough for a long enough period for me to be concerned. Our data mix is thousands of tiff and pdf files under <100K and then lots of fairly large mail folders. rsync with compression and incremental changes worked perfect. Thanks for your help on this. I guess I should unwind and let the Pure Friggin' Magic in rsync do it's stuff! Mike -- redhat-list mailing list unsubscribe mailto:redhat-list-request@xxxxxxxxxx?subject=unsubscribe https://www.redhat.com/mailman/listinfo/redhat-list
