On 18 Dec 2002, Shyamal Prasad wrote:
> "Mike" == Mike Dresser <[EMAIL PROTECTED]> writes: > > Mike> But since there's no rsync server being used in this case, > Mike> it has to read the entire file anyways. > You don't need a server for the incremental updates to happen. rsync > on the client side invokes rsync on the remote side and does an > incremental backup anyway. That is why you need rsh/ssh or something > before you can rsync without the server. At least on a *nix > box. Perhaps wintel changes that....... > > /Shyamal In this case, he's got a mount to the remote machine. He's doing the same thing as doing a rsync a b locally It's going to have to generate c'sums for a, and for b to know what is different, no? As compared to a normal connection rsync a:/a b, where a:/ is told to do the checksums, and the local client handles its own checksums of b and they just compare checksums over the line, sending small blocks as needed. He's going to have it reading all of a, to generate checksums over the line, and then read all of b locally to generate checksums, and then compare in memory. Now, if he put rsyncd.exe from cygwin in the Windows machine and told it to use that, it'd be a lot more efficient. Or do I understand rsync wrong? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]