A while back, I updated my Tumbleweed installation using the KDE Update tool. When I rebooted the machine, it came up with the GRUB command line. My HD is partitioned into 5 partitions. When I use ‘ls’ on the partitions, the first and last are fine, but the middle three show the messages, “No known filesystem detected.”
In GParted via KNOPPIX, I can mount those drives and display their contents. Partition 2 is a swap partition. Partition 3 and 4 are both ext4 partitions. Partition 3, I can see the directory structure, but not any files or subdirectories. In Partition 4, I can see down to the files.
Running e2fsck on partitions 3 and 4 from the Rescue System, I get the results shown in the linked image here: e2fsck results. Everything looks okay to me, but I’m no fsck expert, so if there’s a problem, I wouldn’t know it by looking at those results. The system still won’t boot, and GRUB, ‘ls’ still shows “No known filesystem detected.”
Any suggestions on how I can repair these two partitions?
On one of my systems, I have Tumbleweed and Leap on different partitions, both using “ext4”. The “grub” installed on Leap cannot access the Tumbleweed file system, but the “grub” installed on Tumbleweed can see both partitions. So I make sure that I use the “grub” installed by Tumbleweed.
This is apparently due the Tumbleweed version of “ext4” using newer features that are not supported by the Leap software.
I’m guessing that there is something similar happening on your system.
If you still have the install media available, you can probably use that to boot Tumbleweed. On the boot menu for the install media there should be an option to Boot a linux system. You can probably use that to get into your system.
I tried that using a Tumbleweed install USB stick, and tried both kernel version available (vmlinuz-6.5.9-1-default, and vmlinuz-6.6.1-1-default). Same result. I just boots to the GRUB command line.
I know nothing about the KDE Update tool, but I do know that you do NOT update Tumbleweed, you UPGRADE it. Instead of updating, at a root terminal, you run:
Upgrading is the only supported method of updating TW. Using an install disk to upgrade is fully supported, though rather uncommon. If the latest available version of install disk is not used, it may downgrade more than upgrade. The upgrade process does not format, and does not affect user data.
His screenshot suggests the larger partition 4 likely contains his /home/ filesystem and 3 is /. Knoppix was apparently able to access the full tree on 4. That suggests he should be able to safely fresh install to 3, and work out access to 4 afterward if the installer objects to the condition of 4.
I have the following TW installation on EXT4, which looks much more like his 3 than his 4:
# e2fsck -pvf /dev/sda23
198340 inodes used (38.74%, out of 512000)
59749 non-contiguous files (30.1%)
1104 non-contiguous directories (0.6%)
# of inodes with ind/dind/tind blocks: 50651/2098/3
7107859 blocks used (86.84%, out of 8185084)
0 bad blocks
0 large files
161141 regular files
29296 directories
0 character device files
0 block device files
0 fifos
3640 links
7887 symbolic links (7677 fast symbolic links)
7 sockets
------------
201971 files
#
That is correct. However, I do have a couple of data directories that are not in /home, but rather on the / partition. These are samba share directories (I think that’s why I did not put them in /home). I’d like to recover them if at all possible. One of them is my audio library, and I’d rather not have to start ripping my CD library from scratch again. Ugh. So a fresh install on my 3rd partition is what I’m trying to avoid.
In that case, use a bootable linux device, duplicate those items (either to the external device or to another partition on the drive that won’t be touched by the install).
I would personally cd to the directory where I want to duplicate those items, where as root I would:
cp -R --preserve=all path-to-file-on-/ .
There is a space between that path and the last “.” You probably already know this, but for others who might not: The “-R” is recursive for subdirectories and the “.” means all will be copied to your current directory, which is why I say to “cd to it”.
After you reinstall, just cd to the correct directory and reverse the procedure.
There is a chance that the reinstall will leave the originals intact, but for safety, I would still do the above backup.
I checked again. It turns out that my data, that I thought was on /dev/sda3, must just be symlinks. All of that data is on separate drives. /dev/sda4 is actually /var. I think I have it set up that way from a really old install. I’ve been upgrading the system without changing the partitioning scheme for many, many years.
So, doing a re-install will work, though I’ll have to re-install a bunch of programs. But that’s not that big a deal.
Yeah. I realized that after I replied. When I did realize what I’d said, it was too late to edit the post.
I did the Upgrade, and all three problem partitions are now working fine again. Thanks again for the help. Now I just have to get my Samba shares figured out (but that’s for a different thread).
