I have cloned my master drive so I can use it to do a update to 11.4 from 11.3 preserving my original install. It boots fine then chokes on this looking for the drive by “disk id” during the boot up process. How do I get past this? I have tried both the normal and safe modes.
The disk id contains the manufacturer model ID so it will be different on a new disk. You need to boot with a rescue disk and edit it in two places, the GRUB menu and /etc/fstab. You can choose UUID or device path instead if you want to avoid this kind of issue in future.
I need a blow by blow description on solving this issue.
Editing grub is pretty must past me but I’ll try it if I can cut and paste.
I can manage to edit fstab I think if I can cut and paste the drive id if I can find it somehow.
There is no openSUSE rescue disk that I know of. It’s all gnome **** based on ubuntu.
I have no clue about grub at all except is it hosed you can’t boot up. These are identical model drives.
Why is the default to make a cloned drive fail to boot it seems a poor choice for exactly this reason. The reason I cloned my master drive was to avoid this mess and recover my install if something failed. This stuff should be configured so the default is “just works” for desktop users.
Using Pmagic and I can’t fine the disk id’s I can edit fstab but I can find anything that looks like the id’s in fstab with any of the tools I have tried so far.
How to I choose UID or disk path and where I’ll just re-clone the root directory on my original drive?
On 2011-09-26 01:26, FlameBait wrote:
> There is no openSUSE rescue disk that I know of. It’s all gnome ****
> based on ubuntu.
Yes, there is. The DVD doubles as rescue system, text mode only. Otherwise,
you can use any of the gnome or kde live CDs for any version as rescue system.
–
Cheers / Saludos,
Carlos E. R.
(from 11.4 x86_64 “Celadon” at Telcontar)
Actually I found the choices in the YAST bootloader choices on this machine so it can be done from YAST. My question is now will changing the choices in the bootloader hose up fstab? Or even better is how do I find the disk id’s so I can just edit the files on the cloned drive?
I am googling too. I’d love to not have to do anything to my original drive as it is the back up and shouldn’t be touched.
LOL the drive ID is written right on top of it.:\
Now I need to find out what and where I edit this stuff in GRUB ( like WTH is grub located and what do I need to do to make it play.)
On 2011-09-26 01:56, FlameBait wrote:
>
> Actually I found the choices in the YAST bootloader choices on this
> machine so it can be done from YAST. My question is now will changing
> the choices in the bootloader hose up fstab? Or even better is how do I
> find the disk id’s so I can just edit the files on the cloned drive?
> I am googling too. I’d love to not have to do anything to my original
> drive as it is the back up and shouldn’t be touched.
Common, it is not that difficult.
ls -l /dev/disk/by-*/
–
Cheers / Saludos,
Carlos E. R.
(from 11.4 x86_64 “Celadon” at Telcontar)
Carlos I used that command but I didn’t use the * it returned info I couldn’t use and I googled for disk by id at the command line. It never gave me “ls -l /dev/disk/by-*/” as a result.
I have fstab edited now I have to google the location of the proper file to edit for GRUB to work in openSUSE 11.X.
Carlos I Googled for that command and the ones I tried didn’t give me and useful info. None returned “ls -l /dev/disk/by-*/” as a result. So common for you not common for Google.
I have fstab edited now I have to hope Google can find me the location of grub in openSUSE and the exact file I need to edit and it exact location.
On 2011-09-26 02:46, FlameBait wrote:
>
> Carlos I used that command but I didn’t use the * it returned info I
> couldn’t use and I googled for disk by id at the command line. It never
> gave me “ls -l /dev/disk/by-*/” as a result.
Well, it does. You have by id, by uuid, by path, and by label. You choose
which one you use.
> I have fstab edited now I have to google the location of the proper
> file to edit for GRUB to work in openSUSE 11.X.
Don’t tell me you do not know what file to edit for grub, because I do not
believe you. It comes thousands of times in this forum.
/boot/grub/menu.lst.
You might also have to edit “/boot/grub/device.map”.
–
Cheers / Saludos,
Carlos E. R.
(from 11.4 x86_64 “Celadon” at Telcontar)
Looks like I have a file for every kernel ever installed on my system in /boot what a mess.
Don’t believe me then Carlos. I have no freaking clue. Thanks for the pointers since Google returns a bunch of outdated information with the search terms I used.
The device map file in there points to drives that haven’t been in this system in a long time if I was looking at the right one. Stumbling along here but stumbling forward.
Once I fixed my typo in fstab I have a booted system over on the 11.3 box. I had to edit /boot/grub/menu.list only now I can do the “upgrade”
Sorry I am so clueless and forget everything once I have done it. 
Once I fixed my typo in fstab I have a booted system over on the 11.3 box. I had to edit /boot/grub/menu.list only now I can do the “upgrade”
Sorry I am so clueless and forget everything once I have done it.
I hope you are able to sort this problem, now that you have encountered. In truth, BEFORE cloning IMHO you should have changed your drive calls to be more generic and not by ID.
At the risk of now telling you to CLOSE the barn door, AFTER the horse has gone, if you look in the faq on the use of ‘dd’ for cloning: Cloning an old drive to a new drive you will note it states:
Another caution, if one is cloning a Linux partition, prior to cloning it is likely a good idea to check the /etc/fstab file and the /boot/grub/menu.lst of the old hard drive, to see if it uses disk-by-id identification in the /etc/fstab, the /boot/grub/menu.lst, and the /boot/grub/device.map files. If it does, IMHO it is best to change the fstab, menu.lst, and device.map files to mount/label more generically (ie not by disk-id) and confirm that works prior to cloning. For new users (to Linux) , one can obtain guidance on how to do this by posting a query on this forum. In essence one simply needs to type su -c ‘fdisk -l’ (enter root password) to determine how one’s disks are labeled, and also cat /etc/fstab to see what one’s fstab has for mounting on boot.
For example, prior to cloning assuming fdisk -l gives this:
Device Boot Start End Blocks Id System /dev/sda1 1 3187 25599546 7 HPFS/NTFS /dev/sda2 3188 5099 15358140 b W95 FAT32 /dev/sda3 * 5100 19929 119121975 5 Extended /dev/sda5 5100 8364 26218080 83 Linux /dev/sda6 8364 19721 91233072 83 Linux /dev/sda7 19722 19929 1670728+ 82 Linux swap / Solarisand prior to cloning cat /etc/fstab gives this:
/dev/disk/by-id/ata-Maxtor_6Y160P0_Y4319JDE-part7 swap swap defaults 0 0 /dev/disk/by-id/ata-Maxtor_6Y160P0_Y4319JDE-part5 / ext3 acl,user_xattr 1 1 /dev/disk/by-id/ata-Maxtor_6Y160P0_Y4319JDE-part6 /home ext3 acl,user_xattr 1 2 /dev/disk/by-id/ata-Maxtor_6Y160P0_Y4319JDE-part1 /windows/C ntfs-3g users,gid=users,fmask=133,dmask=022,locale=en_US.UTF-8 0 0 /dev/disk/by-id/ata-Maxtor_6Y160P0_Y4319JDE-part2 /windows/D vfat users,gid=users,umask=0002,utf8=true 0 0then prior to clonging one could change the fstab to this:
/dev/sda7 swap swap defaults 0 0 /dev/sda5 / ext3 acl,user_xattr 1 1 /dev/sda6 /home ext3 acl,user_xattr 1 2 /dev/sda1 /windows/C ntfs-3g users,gid=users,fmask=133,dmask=022,locale=en_US.UTF-8 0 0 /dev/sda2 /windows/D vfat users,gid=users,umask=0002,utf8=true 0 0By changing this prior to cloning (and testing it prior to cloning), one can then not worry about a boot failure after cloning, because of disk specific ids in one’s /etc/fstab file.
A similar edit should be done to the /boot/grub/menu.lst and /boot/grub/device.map. Be very careful here as a bad edit will prevent one’s PC from booting the boot manager.
In my not so humble estimation the default ought to be the simplest form of doing the drive identification ought to be the default.
That doesn’t look like this:
/dev/disk/by-id/ata-WDC_WD5000AADS-00M2B0_WD-WCAV5R429030-part1 / ext4 acl,user_xattr 1 1
/dev/disk/by-id/ata-WDC_WD5000AADS-00M2B0_WD-WCAV5R429030-part2 /home ext4 acl,user_xattr 1 2
proc /proc proc defaults 0 0
sysfs /sys sysfs noauto 0 0
debugfs /sys/kernel/debug debugfs noauto 0 0
usbfs /proc/bus/usb usbfs noauto 0 0
/dev/disk/by-id/ata-WDC_WD5000AAKS-22A7B0_WD-WMASY4550447-part1 /moreAnime ext3 user,acl 1 2
/dev/disk/by-id/ata-WDC_WD5000AVDS-63U7B0_WD-WCAV95924946-part2 swap swap defaults 0 0
/dev/disk/by-id/ata-WDC_WD5000AVDS-63U7B0_WD-WCAV95924946-part1 /bu ext4 acl,user_xattr 1 2
/dev/disk/by-id/ata-WDC_WD5002ABYS-02B1B0_WD-WCASY7880549-part1 /data ext4 acl,user_xattr 1 2
#/dev/disk/by-id/ata-WDC_WD5000AADS-00S9B0_WD-WCAV95847745-part1 /last ext4 acl,user_xattr 1 2
/dev/disk/by-id/ata-WDC_WD5000AADS-56S9B0_WD-WCAV97260650-part1 /mmshare ext4 acl,user_xattr 1 2
It looks like your last case is the best way oldcpu. I am sure the sysadmin types that do this stuff daily will disagree. I use my Linux on the desktop. I’d love to be catered too somewhat by keeping it simple. Imagine if I had mixed manufactures devices what a nightmare that would have been. I was able to wing it off the information off the top of the drive in my case after failing to find the right command to get the device ID exactly as it looks in fstab any place.
IMO, openSUSE should switch to using by-uuid. by-path has the disadvantage that the presence of removable devices can upset the ordering of the devices. by-id has the disadvantage as you have seen of being tied to the model of the disk drive.
On 09/26/2011 09:06 AM, FlameBait wrote:
>
> In my not so humble estimation the default ought to be the simplest form
> of doing the drive identification ought to be the default.
there are advantages and disadvantages to each method of disk
identification…
the advantage of (for example)
“ata-WDC_WD5000AAKS-22A7B0_WD-WMASY4550447-part1” is the system will
‘know’ it no matter where it happens to be placed in the system on the
next boot…like, maybe you install three more harddisks…or . . .
i agree: for the home user (a small subset of the SUSE universe)
/dev/sda1 is much simpler for humans to read.
–
DD
openSUSE®, the “German Automobiles” of operating systems
On 2011-09-26 10:21, DenverD wrote:
> i agree: for the home user (a small subset of the SUSE universe) /dev/sda1
> is much simpler for humans to read.
Except that on some computers what today is, say, sdb, tomorrow is sdf. Not
kidding. Mine does.
–
Cheers / Saludos,
Carlos E. R.
(from 11.4 x86_64 “Celadon” at Telcontar)
If this is so then the instructions on the commands required needed to determine that exact string of characters symbols and numerals needs to be much more explicitly stated. No guessing required kind of instructions in prominent places where those thinking of cloning might avail themselves of them.