Date: Mon, 11 Feb 2002 07:08:14 -0800 From: "Kevin Oberman" <oberman@es.net> To: jacks@sage-american.com Cc: Giorgos Keramidas <keramida@FreeBSD.ORG>, freebsd-questions@FreeBSD.ORG Subject: Re: Using dd to clone HD Message-ID: <20020211150814.E61CB5D0C@ptavv.es.net> In-Reply-To: Your message of "Sun, 10 Feb 2002 07:56:53 CST." <3.0.5.32.20020210075653.0195ca18@mail.sage-american.com>
next in thread | previous in thread | raw e-mail | index | archive | help
> Date: Sun, 10 Feb 2002 07:56:53 -0600 > From: jacks@sage-american.com > Sender: owner-freebsd-questions@FreeBSD.ORG > > Thanks for the reply. I went ahead and tried the 'dd' approach using two > identical 10GB HDs on an experimental box where I wasn't concerned about > the result just to see what would happen. After more than two hours of > copying, I decided to abort the process because 10GB is a pretty small HD > and it would be a very long process to use on the bigger HDs. > > Of course the abort trashed the 2nd HD but fixed it with FDISK. Back to the > drawing board, perhaps with some of your other suggestions. I already use > tar.... I use dd to mirror 2 slices of my hard drive totaling 6 GB. It takes <14 minutes on my UDMA33 laptop disks. 10GB should not take over 4x as long! Do you have write-cache on? This is HUGE for dd. Turn it off and my 14 minute copy turns into a >1 hour copy. The performance decreases by about a factor of 5. Of course, this is very dependent on the exact hardware (controllers, drives) you use. Big block sizes help a lot. I run 32K. 64K would probably be better, if your geometry will allow it. You might also look at the team(1) port. It might allow you to emulate the disk cache in RAM and restore performance without turning on the disk write cache. I have seen reports from others that it is quite effective with dd. Since dd copies every block, used or not, it may work better to use dump/restore for things that are not heavily utilized. But it is far less efficient, so if anywhere near all of the disk in use, dd will run much faster. R. Kevin Oberman, Network Engineer Energy Sciences Network (ESnet) Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab) E-mail: oberman@es.net Phone: +1 510 486-8634 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20020211150814.E61CB5D0C>