My SuSE 8.2 system won't boot, giving the error message: VFS: Cannot open root device "hda3" on 03:03" Booting with the CD or a 7.3 system on another drive, I can mount /dev/hda3 and it looks fine. Everything except the boot partition (hda1) is reiser, and it seems to be finding the boot partition. I tried "boot=/dev/hda3" with the same result. A power failure took the system down. It has been running for over a year and has rebooted successfully many times. I did not change anything recently. (i.e. no new kernel, etc.) The hda3 partition is certainly there, as I can mount it if I boot 7.3 system, but I don't really have any experience with reiser filesystems, so I don't know what to do. I ran fsck.reiserfs (or whatever its name is) off of the intallation CD. It did consistency checks in read-only mode and did not complain about anything. I'm stumped. Help!! -- John Wilkes john at wilkes dot com
I tried "boot=/dev/hda3" with the same result.
Oops, typo. That's "root" not "boot". I should have said that I tried "root=/dev/hda3" as a boot option but it still complains that it cannot open root device on "hda3" on 03:03. To summarize, I can't boot after yesterday's power failure. Same kernel I've been running successfully for a long time. I can boot an old SuSE 7.3 system on /dev/hdb, mount /dev/hda3, and everything looks like it's there. I booted the recovery system on the installation CD and ran reiserfsck, which reported no errors. I'm using GRUB, not LILO, if that matters. I don't really want to re-install everything, but that's what I'm looking at if I can't resolve this today. Any help will be greatly appreciated. /john -- John Wilkes john at wilkes dot com
The Wednesday 2004-04-28 at 07:28 -0700, John Wilkes wrote:
system, but I don't really have any experience with reiser filesystems, so I don't know what to do. I ran fsck.reiserfs (or whatever its name is) off of the intallation CD. It did consistency checks in read-only mode and did not complain about anything.
Try reiserfsck from the rescue system (CD or DVD). Otherwise, check that etc/fstab matches your partition numbers and types. -- Cheers, Carlos Robinson
On Apr 28, 2004, at 4:50 PM, Carlos E. R. wrote:
Try reiserfsck from the rescue system (CD or DVD).
I did. It does the consistency checks and reports no errors.
Otherwise, check that etc/fstab matches your partition numbers and types.
I can mount /dev/hda3 on /mnt from the rescue system and look around. My /mnt/etc/fstab says that /dev/hda3 is mounted on / and is a reiserfs, which is correct. The last few lines of boot output are: Freeing initrd memory: 501k freed VFS: Mounted root (ext2 filesystem). <<== is this the ramdisk? Loading module reiserfs... Using /lib/modules/2.4.20-4GB-athlon/kernel/fs/reiserfs/reiserfs.o kmod: failed to exec /sbin/modprobe -s -k nls_iso8859-1, errno = 2 kmod: failed to exec /sbin/modprobe -s -k nls_iso8859-1, errno = 2 VFS: Cannot open root device "hda3" or 03:03 Please append a correct "root=" boot option You may want to retry boot with pci=noacpi or acpi=oldboot on the command line ... etc. I tried pci=noacpi and acpi=oldboot, and neither one made any difference. I tried adding root=/dev/hda3 to the command line. No difference. Likewise root=0303. What is nls_iso8859-1 and why is boot complaining about it? Guidance and insight will be greatly appreciated. /jw -- John Wilkes john at wilkes dot com
John Wilkes wrote:
On Apr 28, 2004, at 4:50 PM, Carlos E. R. wrote:
Try reiserfsck from the rescue system (CD or DVD).
I did. It does the consistency checks and reports no errors.
Otherwise, check that etc/fstab matches your partition numbers and types.
I can mount /dev/hda3 on /mnt from the rescue system and look around. My /mnt/etc/fstab says that /dev/hda3 is mounted on / and is a reiserfs, which is correct.
The last few lines of boot output are:
Freeing initrd memory: 501k freed VFS: Mounted root (ext2 filesystem). <<== is this the ramdisk? Loading module reiserfs... Using /lib/modules/2.4.20-4GB-athlon/kernel/fs/reiserfs/reiserfs.o kmod: failed to exec /sbin/modprobe -s -k nls_iso8859-1, errno = 2 kmod: failed to exec /sbin/modprobe -s -k nls_iso8859-1, errno = 2 VFS: Cannot open root device "hda3" or 03:03 Please append a correct "root=" boot option You may want to retry boot with pci=noacpi or acpi=oldboot on the command line ... etc.
I tried pci=noacpi and acpi=oldboot, and neither one made any difference. I tried adding root=/dev/hda3 to the command line. No difference. Likewise root=0303.
What is nls_iso8859-1 and why is boot complaining about it?
Guidance and insight will be greatly appreciated.
/jw -- John Wilkes john at wilkes dot com
I think that module is for VFAT support. Check "/etc/fstab" after mounting from rescue boot. HTH -- Louis D. Richards LDR Interactive Technologies
What is nls_iso8859-1 and why is boot complaining about it?
I think that module is for VFAT support. Check "/etc/fstab" after mounting from rescue boot.
No VFAT in my fstab. I am not familiar with details of the boot process, but I infer from the output that /dev/hda3 is never mounted as the root partition, so I am skeptical that the problem could be in any file on that partition. Based on what I've found with Google, it's acting like the kernel does not have reiserfs support, but I don't think that's the case either because the "failsafe" boot, which boots the distributed kernel using the distributed initrd, exhibits the same problem. Maybe I'll try 20 minutes of SuSE support for forty bucks and see if they can get me sorted out. Thanks for the suggestion, and if you or anybody else has any ideas, I'd be most grateful to hear them. /jw -- John Wilkes john at wilkes dot com On Apr 29, 2004, at 6:40 AM, Louis Richards wrote:
John Wilkes wrote:
On Apr 28, 2004, at 4:50 PM, Carlos E. R. wrote:
Try reiserfsck from the rescue system (CD or DVD).
I did. It does the consistency checks and reports no errors.
Otherwise, check that etc/fstab matches your partition numbers and types.
I can mount /dev/hda3 on /mnt from the rescue system and look around. My /mnt/etc/fstab says that /dev/hda3 is mounted on / and is a reiserfs, which is correct.
The last few lines of boot output are:
Freeing initrd memory: 501k freed VFS: Mounted root (ext2 filesystem). <<== is this the ramdisk? Loading module reiserfs... Using /lib/modules/2.4.20-4GB-athlon/kernel/fs/reiserfs/reiserfs.o kmod: failed to exec /sbin/modprobe -s -k nls_iso8859-1, errno = 2 kmod: failed to exec /sbin/modprobe -s -k nls_iso8859-1, errno = 2 VFS: Cannot open root device "hda3" or 03:03 Please append a correct "root=" boot option You may want to retry boot with pci=noacpi or acpi=oldboot on the command line ... etc.
I tried pci=noacpi and acpi=oldboot, and neither one made any difference. I tried adding root=/dev/hda3 to the command line. No difference. Likewise root=0303.
HTH
--
Louis D. Richards LDR Interactive Technologies
-- Check the headers for your unsubscription address For additional commands send e-mail to suse-linux-e-help@suse.com Also check the archives at http://lists.suse.com Please read the FAQs: suse-linux-e-faq@suse.com
On Thursday 29 April 2004 17.16, John Wilkes wrote:
Based on what I've found with Google, it's acting like the kernel does not have reiserfs support, but I don't think that's the case either because the "failsafe" boot, which boots the distributed kernel using the distributed initrd, exhibits the same problem.
Not sure what you mean here, but by default, 'failsafe' boots exactly the same kernel, with exactly the same initrd as the default, only with a few parameters that make it more likely to boot properly in situations where for example acpi makes it fail Are you absolutely certain you haven't had a recent kernel update? There has been some kernel bugs lately. Try creating a new initrd, just for the hell of it and see what happens
The "failsafe" boot uses "kernel.shipped" and "initrd.shipped" on my system. This is how the standard SuSE 8.2 installation set up grub's menu.lst file. I re-installed the kernel off of the distribution CDs, which gets me a new initrd. No difference. Still complains that it can't open the root device. I'm thinking about backing up the partition and nuking it with mkreiserfs, then restoring its contents. Is tar suitable for this purpose? I've heard in the past that tar is not good for backups of this sort. Thanks for the suggestions. -- John Wilkes john at wilkes dot com On Fri, 30 Apr 2004, Anders Johansson wrote:
On Thursday 29 April 2004 17.16, John Wilkes wrote:
Based on what I've found with Google, it's acting like the kernel does not have reiserfs support, but I don't think that's the case either because the "failsafe" boot, which boots the distributed kernel using the distributed initrd, exhibits the same problem.
Not sure what you mean here, but by default, 'failsafe' boots exactly the same kernel, with exactly the same initrd as the default, only with a few parameters that make it more likely to boot properly in situations where for example acpi makes it fail
Are you absolutely certain you haven't had a recent kernel update? There has been some kernel bugs lately. Try creating a new initrd, just for the hell of it and see what happens
The Sunday 2004-05-02 at 22:15 -0700, John Wilkes wrote:
I'm thinking about backing up the partition and nuking it with mkreiserfs, then restoring its contents. Is tar suitable for this purpose? I've heard in the past that tar is not good for backups of this sort.
I use mc (midnight comander) myself, I'm lazy. For a complete howto, see: /usr/share/doc/howto/en/mini/Hard-Disk-Upgrade.gz -- Cheers, Carlos Robinson
(Sorry for the delayed response.) My hda disk has SuSE 8.2 on it. That's the one that can't open hda3 as the root partition. The hdb disk has an old SuSE 7.3 system on it. Running 7.3, I can mount /dev/hda3 and /dev/hda1. /etc/fstab: /dev/hda3 / reiserfs defaults 1 1 /dev/hda1 /boot ext2 defaults 1 2 /dev/hdb1 /data1 auto noauto,user 0 0 /dev/hdb3 /data2 auto noauto,user 0 0 /dev/hdb5 /data3 auto noauto,user 0 0 /dev/hdb6 /data4 auto noauto,user 0 0 /dev/hdb7 /data5 auto noauto,user 0 0 /dev/hda8 /home reiserfs defaults 1 2 /dev/hdb8 /home2 ext2 noauto,user 1 2 /dev/hda6 /opt reiserfs defaults 1 2 /dev/hda5 /usr reiserfs defaults 1 2 /dev/hda7 /var reiserfs defaults 1 2 /dev/hda2 swap swap pri=42 0 0 /dev/hdb2 swap swap pri=42 0 0 devpts /dev/pts devpts mode=0620,gid=5 0 0 proc /proc proc defaults 0 0 usbdevfs /proc/bus/usb usbdevfs noauto 0 0 /dev/cdrecorder /media/cdrecorder auto ro,noauto,user,exec 0 0 /dev/cdrom /media/cdrom auto ro,noauto,user,exec 0 0 /dev/dvd /media/dvd auto ro,noauto,user,exec 0 0 I re-installed the kernel from the 8.2 installation CD, which also gave me a new initrd. These fail in the same way. /boot/grub/menu.lst: # Modified by YaST2. Last modification on Thu Apr 29 13:52:56 2004 color white/blue black/light-gray default 0 gfxmenu (hd0,0)/message timeout 8 title linux kernel (hd0,0)/vmlinuz root=/dev/hda3 vga=0x31a hdc=ide-scsi hdclun=0 splash=silent showopts initrd (hd0,0)/initrd title SuSE_7.3 kernel (hd1,0)/vmlinuz root=/dev/hdb3 title failsafe kernel (hd0,0)/vmlinuz.shipped root=/dev/hda3 showopts ide=nodma apm=off acpi=off vga=normal nosmp noapic maxcpus=0 3 initrd (hd0,0)/initrd.shipped title floppy root (fd0) chainloader +1 -- inet: john@wilkes.com | Fascism conceives of the State as an addr: 321 High School Rd. NE #367 | absolute, in comparison with which all city: Bainbridge Island, Washington | individuals or groups are relative, code: 98110-1697 | only to be conceived of in their icbm: 47 37 48 N / 122 29 52 W | relation to the State. - B. Mussolini On Fri, 30 Apr 2004, Carlos E. R. wrote:
The Thursday 2004-04-29 at 00:03 -0700, John Wilkes wrote:
Guidance and insight will be greatly appreciated.
Could you post your /etc/fstab file? And perhaps your /boot/grub/menu.lst as well.
The Sunday 2004-05-02 at 22:08 -0700, John Wilkes wrote:
(Sorry for the delayed response.)
No problem.
My hda disk has SuSE 8.2 on it. That's the one that can't open hda3 as the root partition. The hdb disk has an old SuSE 7.3 system on it. Running 7.3, I can mount /dev/hda3 and /dev/hda1.
/etc/fstab:
/dev/hda3 / reiserfs defaults 1 1 /dev/hda1 /boot ext2 defaults 1 2
/dev/hdb1 /data1 auto noauto,user 0 0 /dev/hdb3 /data2 auto noauto,user 0 0 /dev/hdb5 /data3 auto noauto,user 0 0 /dev/hdb6 /data4 auto noauto,user 0 0 /dev/hdb7 /data5 auto noauto,user 0 0
/dev/hda8 /home reiserfs defaults 1 2
/dev/hdb8 /home2 ext2 noauto,user 1 2
/dev/hda6 /opt reiserfs defaults 1 2 /dev/hda5 /usr reiserfs defaults 1 2 /dev/hda7 /var reiserfs defaults 1 2
/dev/hda2 swap swap pri=42 0 0 /dev/hdb2 swap swap pri=42 0 0
devpts /dev/pts devpts mode=0620,gid=5 0 0 proc /proc proc defaults 0 0 usbdevfs /proc/bus/usb usbdevfs noauto 0 0 /dev/cdrecorder /media/cdrecorder auto ro,noauto,user,exec 0 0 /dev/cdrom /media/cdrom auto ro,noauto,user,exec 0 0 /dev/dvd /media/dvd auto ro,noauto,user,exec 0 0
As far as I see, it is correct.
I re-installed the kernel from the 8.2 installation CD, which also gave me a new initrd. These fail in the same way.
See if file "/etc/sysconfig/kernel" contains the line: INITRD_MODULES="reiserfs" If not, add it and run "mk_initrd".
/boot/grub/menu.lst:
# Modified by YaST2. Last modification on Thu Apr 29 13:52:56 2004
color white/blue black/light-gray default 0 gfxmenu (hd0,0)/message timeout 8
title linux kernel (hd0,0)/vmlinuz root=/dev/hda3 vga=0x31a hdc=ide-scsi hdclun=0 splash=silent showopts initrd (hd0,0)/initrd
That should be booting "vmlinuz" in the first IDE disk, first partition. Ie, hda1 should be "/boot". That is correct, I think. however, you reported the error as:
VFS: Cannot open root device "hda3" on 03:03"
That, according to grub manual, should be hardisk number four (it is cero based), and partition number four (primary) - that doesn't make much sense, if the error comes from grub. Have you changed the ordering of disks in the bios? Boot sequence? -- Cheers, Carlos Robinson
On Monday 03 May 2004 13.00, Carlos E. R. wrote:
VFS: Cannot open root device "hda3" on 03:03"
That, according to grub manual, should be hardisk number four (it is cero based), and partition number four (primary) - that doesn't make much sense, if the error comes from grub.
That message comes from the kernel and refers to the major/minor number of the device, and hda3 is major 3, minor 3, which is consistent with the other info
On Mon, 3 May 2004, Carlos E. R. wrote:
See if file "/etc/sysconfig/kernel" contains the line:
INITRD_MODULES="reiserfs"
It does.
That should be booting "vmlinuz" in the first IDE disk, first partition. Ie, hda1 should be "/boot". That is correct, I think.
Correct.
however, you reported the error as:
VFS: Cannot open root device "hda3" on 03:03"
That, according to grub manual, should be hardisk number four (it is cero based), and partition number four (primary) - that doesn't make much sense, if the error comes from grub.
You are correct that grub numbers from zero, but that message is coming from the kernel on the /boot partition, not grub.
Have you changed the ordering of disks in the bios? Boot sequence?
No. I can boot the SuSE 7.3 system I have on my /dev/hdb disk and mount /dev/hda3 on /mnt. I don't need that old 7.3 system, and I intend to install SuSE 8.2 on top of it to get mostly back to where I was. I've mounted the /var partition and looked at the /var/log/messages file, but there is nothing suspicious in there. Specifically, no disk I/O errors reported. /jw -- John Wilkes john at wilkes dot com
The Monday 2004-05-03 at 07:34 -0700, John Wilkes wrote: ...
however, you reported the error as:
VFS: Cannot open root device "hda3" on 03:03"
That, according to grub manual, should be hardisk number four (it is cero based), and partition number four (primary) - that doesn't make much sense, if the error comes from grub.
You are correct that grub numbers from zero, but that message is coming from the kernel on the /boot partition, not grub.
Ah, I thought that could be so, I got confused.
Have you changed the ordering of disks in the bios? Boot sequence?
No.
Pity... everything seems correct, but something must not be so - we are not progressing much. Memory errors? Once I got a corruptions because I had to reseat my memory chip. Run the memory check program from grub.
I can boot the SuSE 7.3 system I have on my /dev/hdb disk and mount /dev/hda3 on /mnt. I don't need that old 7.3 system, and I intend to install SuSE 8.2 on top of it to get mostly back to where I was.
I also keep two systems on the same machine, one 8.2 and the emergency one is 7.3. Same thing.
I've mounted the /var partition and looked at the /var/log/messages file, but there is nothing suspicious in there. Specifically, no disk I/O errors reported.
There wouldn't show much, because that partition is not mounted at the time of your problems. What are the messages you see before the last one (VFS...)? There might be a clue. Did you say the failsafe option from grub worked? If you have reinstalled the kernel, both vmlinuz.shipped and vmlinuz are identical. In that case, it must be a boot option it doesn't like... Copy vmlinuz.shipped over vmlinuz, and initrd.shipped over initrd Another idea: Use the failsafe boot, and remove options till it stops booting. -- Cheers, Carlos Robinson
On Tue, 4 May 2004, Carlos E. R. wrote:
Did you say the failsafe option from grub worked?
No. I said the failsafe option did not work. The 7.2 system on hdb works fine. (Using it now.) So I doubt memory errors are the problem, but I'll run the memtest from grub just to be sure. Thanks for the suggestions. /jw -- John Wilkes john at wilkes dot com
participants (4)
-
Anders Johansson -
Carlos E. R. -
John Wilkes -
Louis Richards