On Mon, 2003-07-28 at 18:22, Alan Connor wrote: > There have been endless discussions about this on various linux groups, and > the consensus is that dd is not a good idea for this.
dd has never failed me. My situation is different than his, but I've used it extensively. I have a Thinkpad, and there's no cost effective way to backup up 20G. I bought a second, identical drive and a carrier to mount it in the extra drive bay, and I have backed up and restored my system for nearly two years with this method. Another advantage is I can restore the backup in the 2 minutes it takes to physically swap the drives, and redo the backup later. My Dad now has 2 40G drives in his desktop - how much hassle and expense is a 40G tape backup? Or 2 or 3 5G tapes? Using maybe two or more backup schemes/programs for his two OSs? I don't know the result if I used a target that's larger or smaller. If I specify at the device level (/dev/hda) then I'm copying all the mbr and partition info. When I specify at the partition level (/dev/hda1), I'm copying bytes and it might work to transfer to larger or smaller partitions. I'd think I'd have to format the target at least to expect it to work afterwards, and in that case, I'd think "cp -a[u]" would be the way to go myself. But in his case, with distinct drives filtered through a raid card, I'll admit I don't know and I'm curious. If the aggregate sizes match (or exceed in the right direction), it might work. If he can afford the electricity, and the time to do it twice, it may be worth the gamble. I don't see any way he can damage either source or target, and if it works, he's learned something interesting as well as saved some hassle. Cheers, Bret -- bwaldow at alum.mit.edu -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
