NavigationUser loginSpam?See spam posts on this site? If so, please don't reply to the spam! Instead, just report the URL to the webmaster. Cost of War |
Etch upgrade - /dev/hda1 disappearing.I recently did a dist-upgrade of my Sarge system to etch. Now, I followed the guidelines in the upgrade suggestions, and did the install of initrd-tools and so on. Still, when I decided to restart, the kernel (2.6.18-4, I believe) gave the following message (after hanging for a minute or so): ALERT! /dev/hda1 does not exist And fell to a busybox shell. If I try and fallback to earlier kernels (.14, .8), it just kernel panics instead saying that it can't find the root partition. I expect that this is probably the same issue, just handled worse? (In fairness, this would happen before with the very old kernels too, so it's possible the problem is different? Of course, now the previous kernel which worked fine before now dies to the kernel panic, while the most recent one installed in the upgrade gives the Alert! error). I do have an etch install CD handy, and I've booted that into rescue mode. It certainly seems able to at least mount the root partition, as I recognize some stuff in /root. I can also overwrite initrd in / (which is suspiciously small, I'll add- initrd is 4096 bytes? I assume that the real one, initrd.img, is larger, but that's a symbolic link to the boot directory, which wasn't mounted by the rescue shell.) This problem seems to have occurred before; https://bugs.launchpad.net/ubuntu/+source/initramfs-tools/+bug/32123 for Ubuntu seems to be basically the same issue (makes sense, since Ubuntu is based off Debian). In any case, does anyone have some ideas? I suppose I could try and stick a new kernel image in, but I wouldn't even know what to change. I'd hope there'd be some kind of simple configuration change to make, but I'm not sure where to start, and I can't seem to cheat my way into my pre-reboot setup from the rescue shell. Worst comes to absolute worst, I suppose I could install a new hard drive, put a fresh install of etch there, then mount the old hard drive as a slave to /oldhd or something (and use chroot). I'd hope that there's a better way around it, though. Thanks in advance. |
Etch upgrade - /dev/hda1
Use the busybox to try to find out what went wrong. Obviously the boot from initramfs/initrd was fine but the once-successful /etc/fstab entries now appear incorrect. One thing that happens is 'hda' becomes 'sda' in the new setup, but this doesn't explain why your older kernels are having problems. So, while in the busybox you can try to find out what your partitions are called. I can't remember if there is an 'fdisk' in there, but that would be the easiest:
fdisk -l /dev/somedevice
otherwise you can try:
less -f /dev/somedevice
and you'll get "device does not exist" if 'hda', 'hdb' etc aren't really devices. I prefer 'fdisk' if it's available because it will tell you the partition scheme if the device exists.
Etch upgrade - /dev/hda1
Found the problem (after manually mounting stuff succesfully). I'd updated fstab, but hadn't updated /boot/grub/menu.lst , so it was still passing the wrong argument with an hda there.
Now, all I have to do is figure out why X.org refuses to load and startx dies on running, but that's not exactly a kernel issue... thanks for the help.
reinstall the xorg packages, and the fonts
Just like the title says.
similar problem due to device ordering
I had a similar problem on upgrade from Sarge to Etch. Turned out it was actually the upgrade from kernel 2.6.8 to 2.6.18. Something changes in how the kernel counts devices. I have an unused PATA controller in a PCI slot. 2.6.8 counted it after the motherboard PATA channels. 2.6.18 counts it first. So my boot partition changed from /dev/hda1 to hde1.
Solution is to use volume labels in /etc/fstab and /boot/grub/menu.lst, not device names. Use e2label(8) to add labels to file systems that were created without them. Syntax in fstab is
LABEL=ata-root / ext3 defaults,errors=remount-ro 0 1