I get this error message in a duol boot system (Leap 15.1 and Windows 10)
error: file /boot/grub2/x86_64-efi’ not found
error: file /’ not found
Entering rescue mode…
grub rescue>
I’ve no idea on how to use rescue mode.
Using SuperGrub disk I’m able to start Window 10 but no way to start openSUSE.
This happens after restoring a disk image using Clonezilla in 16 pc because I would avoid to reinstall manually openSUSE 16 times.
What exactly did you “restore” or copy to 16 PCs?
Was it one partition, or several partitions?
You write about a disk image, but has that been complete?
As a first step it would be good if you could boot one of the 16 PCs using a Linux from an external device like a USB or external hard disk.
That could as well be an openSUSE 15.1 installed to a USB.
Then, in a terminal, as root, enter
parted -l
and please post the output here. That output can even be redirected to a file when you have booted a Linux for which you have write access (like an openSUSE 15.1 installed to a USB). The command would then be something like
parted -l >output_parted_l.txt
by which the file output_parted_l.txt is created in your current directory with the output of “parted -l” in it.
Further you could try to mount the partition in which you think that openSUSE Leap 15.1 should exist (that would be /), and check what you find there, for example if there is a directory /boot and others.
I see that Conezilla has a page that describes what you need to do, I didn’t read it closely but assume it should work for you. If you have problems with this page, post again.
TSU
I’ve 16 pcs with same hardware and I want avoid to install 16 times openSUSE. Then when the first pc is complete (operating systems, software, settings …) I made a disk image of this first pc with Clonezilla and I use the image to “copy” all this in other 15 pcs.
Was it one partition, or several partitions?
Several partition, they should be 8.
You write about a disk image, but has that been complete?
I think it’s complete, Clonezilla prompt when image creation is finished.
Here’s what you asked for. JetFlash Transcend 4GB is usb containing live.
Model: ATA INTEL SSDSCKKW51 (scsi)
Disk /dev/sda: 512GB
Sector size (logical/physical): 512B/512B
Partition Table: gpt
Disk Flags:
Number Start End Size File system Name Flags
1 1049kB 556MB 555MB ntfs Basic data partition hidden, diag
2 556MB 661MB 105MB fat32 EFI system partition boot, esp
3 661MB 677MB 16.8MB Microsoft reserved partition msftres
4 677MB 199GB 199GB ntfs Basic data partition msftdata
5 199GB 336GB 136GB ntfs Basic data partition msftdata
6 336GB 389GB 53.7GB btrfs
7 389GB 391GB 2147MB ext4
8 391GB 413GB 21.5GB ext4
Model: JetFlash Transcend 4GB (scsi)
Disk /dev/sdb: 3932MB
Sector size (logical/physical): 512B/512B
Partition Table: msdos
Disk Flags:
Number Start End Size Type File system Flags
1 512B 2618MB 2618MB primary boot
2 2618MB 2625MB 6953kB primary fat16 esp
I understood all this already from your first posting.
I have a question: as it seems you have been booting from a live Linux, so I guess that you have been getting this output on one of the other 15 PCs.
This is important, because the first PC from which you drew the image should boot openSUSE without any problem, so you have to check for one of the 15 PCs that don’t boot openSUSE.
If you would have produced this output on the first PC, then please post that output again, but then obtained on one of the other 15 PCs.
In addition, in order to see how the partitions are attempted to be mounted booting openSUSE, please boot one of the 15 PCs using your live USB and post the contents of /etc/fstab that should be present on partition no 6.
Please remember which one of those 15 PCs that was. You should always work with the same PC that won’t boot openSUSE.
And if possible, please post this in code tags, i.e. when you write your next posting you paste the contents, then mark the pasted text with the mouse cursor, and then you click on ‘#’ above the text window in the web page.
In the mpeg video on
https://forums.opensuse.org/showthread.php/451526-Posting-in-Code-Tags-A-Guide
you’ll find a slightly different way how to do that - the links to the images on that page however don’t work anymore.
Right. I run “parted -l” in one of the “not working pc”.
This is important, because the first PC from which you drew the image should boot openSUSE without any problem, so you have to check for one of the 15 PCs that don’t boot openSUSE.
I made the image of the first pc after all software and settings are complete and working.
In addition, in order to see how the partitions are attempted to be mounted booting openSUSE, please boot one of the 15 PCs using your live USB and post the contents of /etc/fstab that should be present on partition no 6.
Here’s I see one more information. I checked partitions in “not working” pc and they seem correct in size and free space. Using live I can mount “home” and “swap” partitions but I’m not able to mount “root” partition because I get this error:
An error occurred while accessing 'op_SU_Leap_15_1_root', the system responded: The requested operation has failed: Error mounting /dev/sda6 at /media/live/op_SU_Leap_15_1_root: wrong fs type, bad option, bad superblock on /dev/sda6, missing codepage or helper program, or other error
Then I suspect brtfs partition isn’t working. I downloaded an updated version of Clonezilla and will try again to made the image and restore it.
Please remember which one of those 15 PCs that was. You should always work with the same PC that won’t boot openSUSE.
Sure.
OK, I’m probably not able to help you much with that anymore, because I use ext4 for both, / (or root) and /home.
I saw it in the output of “parted -l” that you posted that you have been installing using the default for root, e.g. btrfs.
However you changed the default for /home from xfs to ext4.
Have you changed other things like mount point options?
If in the mount point options (in the partitioner) you would have chosen to mount by ID, then that could not work, because the ID changes with hardware (the hard disk / ssd, and you have 16 of them, not one).
If you would have left it at the default of “Mount by UUID”, then one should check if the right UUID is used in e.g. /etc/fstab or for GRUB2-EFI.
I don’t know if clonezilla changes UUIDs.
It may have been possible to chroot into the installed openSUSE as described in
https://doc.opensuse.org/documentation/leap/startup/html/book.opensuse.startup/cha.trouble.html#sec.trouble.data.recover.rescue.access
but with btrfs I just don’t know if that would work, especially given the error that you encountered.
Then it would have been possible to run the text version of YaST entering the command “yast” at the root / rescue prompt, which would have made it possible to rearrange mount options using the partitioner of yast, and to try to fix GRUB2-EFI using bootloader of yast.
However with the error that you get …
By the way, I just saw this
https://doc.opensuse.org/documentation/leap/startup/html/book.opensuse.startup/cha.trouble.html#sec.trouble.boot.btrfs
while scanning through that page.
Don’t know if that helps.
I encourage you again to use the Clonezilla documentation I provided as a guide.
You also have not provided essential information in your posts about what is in your clone image beyond the 8 partitions, and whether you’re restoring the entire image or only parts of it.
The simplest solution that should work is if you do a Full Disk image because there is almost no room for any kind of errors like what you describe… If the original image works, then the restored image should be an exact duplicate in all ways. If you’re restoring individual partitions, then anything is possible and errors can happen. Note a full disk image will include everything on the disk regardless of number of partitions, filesystem formats, configuration, or anything else… Just whatever bytes are written to the disk.
Keep in mind that restoring a full image to multiple machines means that each machine would be configured with exact same machine identifiers and network settings which will cause varying amounts of problems if active on the same LAN right away… If they’re configured with DHCP settings, then hopefully each will be assigned a different IP address because of the MAC addresses burned in silicon… But otherwise they will have the same machine names, user accounts, services, etc. which if exposed on your LAN would cause conflicts. In fact, there is probably a “Best Practices” guide for creating the golden image containing things like emptying trash, clearing tmp locations and files, removing unnecessary Users, zeroing out unused file system space, if names are unimportant then randomizing on first boot, etc. More intelligent cloning systems provide a way to automatically change these settings for what you are doing but I don’t see a Clonezilla feature that addresses this.
HTH,
TSU
https://forums.opensuse.org/showthread.php/536143-Using-Code-Tags-Around-Your-Paste should be sufficient instruction.
A small comment because I don’t know your cloning setup…
The Clonezilla documentation I described involves either pulling drives and installing in the cloning system or plugging USB keys into all your available ports.
If you’re trying to clone systems directly without pulling drives or inserting dongles,
You might want to look at a PXE boot system…
The idea is to boot every target machine with a client (ie USB key or CD, can be anything that’s bootable),
Only enough is loaded to reach out on your network to the server holding your image and then that image is written to your machine.
My personal favorite for this or any other kind of Enterprise backup/restore is g4u
http://www.feyrer.de/g4u/
Although it used to be somewhat complex due to supporting a large number of transfer protocols and different kinds of images, the last time I looked at this it’s been boiled down to some reasonable basics so it’s pretty straightforward to set up.
TSU
I’ll add a note about “btrfs”.
At one time, I installed openSUSE to boot with 32-bit UEFI. And the normal installer does not support that, so I had to improvise.
After installing, and booted to an external drive, I tried to install grub2. I had mounted the newly installed system, including the “btrfs” subvolumes. So I used something like:
grub2-install --boot-directory=/mnt/boot --efi-directory=/mnt/boot/efi
And the system would not boot. The “grub2-install” said it was successful, but it failed to include support for “btrfs”.
So I then went into chroot rescue mode, and did the grub2-install from there. And that worked.
I don’t know whether that is the problem you are having, but “btrfs” support for grub2 still seems a little tricky.
I much prefer dd to any other clone method
It’s never let me down
Many thanks to all that answered.
The problem was Clonezilla I used: too old to manage btrfs. I suspected that btrfs partition restored was wrong because I can’t access it using live then I downloaded and used last version of Colnezilla (2.6.3-7 at present) and now image restored works flawlessly.
I amnot a Clonzilla user and this will not encourage me to use it. A clone should be a clone. No matter what the bytes represent to other software. Thus there should not be any “knowledge” about Btrfs inside Clonezilla.
For cloning I am with caf4926. dd clones perfectly and does so already for about 50 years.