In Yast’s partitioner>Config you can change the standard mount type from device ID to device name (e.g., /sda1, etc.), device label (volume label) or device path (pci-0000:00:11, etc.), besides UUID.
I’d suggest you make a backup of the three or four files mentioned earlier, change to device name (or label) and swap the drives, making sure you plug the new one in the same IDE/SATA port.
If things go wrong you just swap the drives back.
Worst case scenario, you boot with partition magic (or any other livecd) and restore the three/four files, reboot.
Thanks for this - I sort of hate doing command-line or system-file-edit type changes I’m unfamiliar with because there is usually something involved only a Real Programmer would know, so thanks for the YaST link. I was thinking about this and realized that after I edit the 4 files, I can’t reuse that drive without booting into, say, Knoppix, and fixing it. I keep an old copy of Knoppix around (kde3 - I wonder how many more years it will take me to get facile enough with kde4 to change to that for system fixing?) for this.
I’m looking at the Partitioner right now - is that in the Fstab options button in the Edit… dialog of the Partitions listing under new-host/Hard Disks/sdxxx? Does anyone know if this change gets propagated intelligently throughout the system? (i.e., the 4 files and all dependencies) If so, and if I were worried about future software incompatibilities popping up, then I could make your suggested change, clone the disk, then change back via YaST It does seem to me that the kernel maintainers have settled on by-id disk specification.
Changing the UUID with an ext4 utility should easily solve some problems. Don’t want all your hard drives with the same UUID if cloning fixed and portable drives.
> Does anyone know if this change gets
> propagated intelligently throughout the system? (i.e., the 4 files and
> all dependencies)
Certainly not.
The yast partitioner only changes fstab and partitions.
> It does seem to me that
> the kernel maintainers have settled on by-id disk specification.
Kernel maintainers offer you 4 types of partition identification, without
counting the old /dev/sd*. Each type has its own advantages/disadvantages.
And the Yast partitioner lets you choose the default one to use (for this
session only), and to change the method for each partition.
–
Cheers / Saludos,
Carlos E. R.
(from 11.2 x86_64 “Emerald” at Telcontar)
On 2010-11-07 14:36, ken yap wrote:
>
> Strictly speaking it’s udevd, a userspace program that populates the
> /dev/disk/by-* directories with symlinks.
True. But with information given by the kernel, which obtains it from disks
and hardware.
Now I wonder where that information is. /proc?
–
Cheers / Saludos,
Carlos E. R.
(from 11.2 x86_64 “Emerald” at Telcontar)
Since any non-trivial app has to interact with the kernel in some way, you would have to have a good argument to claim special status for udevd. It’s just a normal executable. Doesn’t even need suid because it’s run by root as part of the init process.
A quick strings on the binary shows that it probably reads synthetic filesystems like /sys and so forth.
You could, but why not do it before you try to clone. The same applies now wrt editing the files back if you have the wrong settings (i.e. if wrong settings before cloning discovered during a pre-clone test, just boot to a liveCD, mount the partitions, and change the file names back … just be certain to have a backup copy that is easy to put back).
Thanks!! My thinking is that if I don’t mess with the bootable drive which I’m cloning, I can always boot to a functioning system. It’s just a backup drive I’m creating, although when my main drive finally does crash, I can just swap drives to the new (cloned) one. If I change the drive-id numbers in those 4 files before cloning, then the drive I’m cloning from (my main drive) is instantly unbootable and needs to be edited (those 4 files) to make it bootable again, right?
If you use the SDXX syntax and change the boot order or cabling of the drives when drive 1 fails you would be back in buisness. Note that this should be done in Yast so kernel updates don’t get confused when rewriting the files and use the wrong syntax.
Oh, OldCPU - I meant to ask you how you’re doing on the Hi-Def thing? I recently found out that my Vixia is progressively-scanned, so moving objects come out “bent.” Also, I had to switch to 11.3x64 on one of my machines because KDEnlive seems to work better there than under 11.1x64. So far, KDEnlive seems to be the best combination of power/ease so that I can actually do full-on editing. Cinelerra probably would also, but it’s too hard for a noob to use. My poor quad-core Phenom 3GB machine struggles with the size of Hi-Def (1920x1200) video files. I think I mainly need more memory. I am not sure if video cards make much of a difference to nonlinear video editing - I think as long as you have a capable GPU, you’re fine - the main delay is surfing through a huge file.
Thank You!!! Actually, I think you’re right, and I was considering doing that because it seemed to be a simple roll-back to the simpler pre-by-id disk days, but some of the messages in this thread seemed to indicate that changing to sdXX-style syntax in YaST might actually break something. Maybe not right away, but after the clone - because YaST leaves some system or app files in the default by-id addressing mode. I’ve seen this kind of problem before. Linux is really complicated, and it would be best to get a Wizard’s ruling on doing it before hosing my system.
Patti
EDIT: Oh, wait a minute, you’re a Shaman - is that like a Wizard](wizard from FOLDOC)? Or is it more like a Lord High Fixer (LOL - I love the FOLDOC stuff…)
“We need to find a wheel to unwedge the hung tape drives.”
> Thank You!!! Actually, I think you’re right, and I was considering
> doing that because it seemed to be a simple roll-back to the simpler
> pre-by-id disk days, but some of the messages in this thread seemed to
> indicate that changing to sdXX-style syntax in YaST might actually break
> something.
And it does.
My system is unbootable some days if I do so. Some days it boots, some not,
because what today is sda another day can be sdf, so that it is not found.
There are four types of IDs, just choose the one that works best in your
intended usage.
–
Cheers / Saludos,
Carlos E. R.
(from 11.2 x86_64 “Emerald” at Telcontar)
appologies if this is not an appropriete question for this thread!
as you can see im a puzzled penguin here!!, i am also keen to clone my hard drive for back up purpose, and at present run
11.3 with windows xp on an 80gb hd, i have just recieved a new 80gb hard drive today,and as it is new sealed item,how
should i format it for the clone…is ntfs ok??..using xp to format!..advice appreciated,cheers J.
I use kdenlive extensively. I have it running on all my PCs (from my lowliest Athlon-1100 with 1.0 GBytes of RAM). But it runs best on my Intel Core i7 920 with 6GB of RAM.
As for playing back Hi-Def (1920x1200) video files, my Intel Core i7 920 is my only PC that plays them back effortlessly. Dependending on the bit rate of the 1920x1200 video file, my 32-bit athlon-2800 with nVidia 8400GS graphics using VDPAU can sometimes play back those 1920x1200 video files.
For processing those files in an editor such as KDEnlive, I typically reduce them to 1280x720.
On 2010-11-09 13:36, diamond1902 wrote:
>
> appologies if this is not an appropriete question for this thread!
> as you can see im a puzzled penguin here!!, i am also keen to clone my
> hard drive for back up purpose, and at present run
> 11.3 with windows xp on an 80gb hd, i have just recieved a new 80gb
> hard drive today,and as it is new sealed item,how
> should i format it for the clone…is ntfs ok??..using xp to
> format!..advice appreciated,cheers J.
A clone is an exact copy, you don’t format it.
A byte by byte copy means even the format is copied.
However, some tools, like ghost (windows) can duplicate a system changing
the filesystem type on the fly, I understand.
–
Cheers / Saludos,
Carlos E. R.
(from 11.2 x86_64 “Emerald” at Telcontar)
sorry! was not making myself clear…i placed new hard drive in pc as slave and it wasnt recognised due to being new and unformatted,i have since formatted it and it is now recognised as sdb, if im correct all i need to do now is…
( please correct me if im wrong !! ) I unmount /dev/sdb…then in terminal as SU i enter dd if=/dev/sda of=/dev/sdb
do i need to add anything to this command or just sit back and allow it to copy ???
It does seem to me that the kernel maintainers have settled on by-id disk specification.
