I was following Carl Spitzer's advice concerning getting my firewire drive to let me write to it(It worked,thanks Carl). He also noticed that in my fstab I had my floppy and dvdrecorder devices under subfs and he suggested I add 'users' to the fs= part of the media under subfs,which I did using vim. But I still couldn't write to the floppy even though I owned it. So I decided to run SuSEconfig-no help. Then I decided to reboot and this has led to a serious problem for me. When I reboot I get the message that: ' fsck failed, my / partition is being mounted read-only' and I must login as root. However when I run mount I get this: (none):~#mount /dev/hdb2 on / type(reseiserfs) (rw) /dev/hdb2 on / type(reseiserfs) (rw) Two entries for the same file system! And the filesystem is mounted (rw). Running: (none):~#fsck fails and I'm told to run it manually. But when I run fsck.reseiserfs --fix-fixable I get this: (none):~#fsck.reseiserfs --fix-fixable Partition /dev/hdb2 is mounted with write permissions, cannot check it. When I mount the system read-only it gets checked but nothing gets fixed. The first time this happened four days ago I was able to solve it by checking the 'initialize the / partition' box in the bootloader configuration dialogue in the rescue script of the 9.1 installation disk. But this time that doesn't work. I've also tried reloading a generic copy of the boot loader. When I try using the advanced setting in the rescue script to initialize my / partition I alway get an error ( something about high memory and low memory). If I tell it to try again the module crashes and I have to reboot. If I use the automatic option I can edit the bootloader module successfully but I have to go the the whole time consuming process. I've been trying to add a copy of my 'boot_problems' file which contains a copy of fstab to a floppy to transfer to this machine. '(none):~#cat /media/floppy/boot_problems' on my unbootable machine shows my boot_problems file but when I mount that floppy on this machine it is empty which I would expect the latter since not having write access to the floppy is why I rebooted in the first place. But I didn't expect the unbootable machine to show me my file when I run 'cat /media/floppy/boot_problems' even when there is no floppy loaded in the unbootable machine.! I don't know what else to do except: (none):~#fsck.reseiserfs --rebuild-tree. But I don't know if this will make matters better or worse or even work since fsck.reseirfs --fix-fixable does not. I've run the automatic rescue script option on the installation disk successfully several times so I'm not convinced its a file system or a boot loader problem. Can anyone help me get my system back? Thanks, Jerome ps At the fsck failed login prompt I log in as root, enter init5 and restartx. I get my GUI back but I don't have access to the web, I get the 'unknown host error' message. Also the prompt is 'user@(none):~> Instead of 'user@(Mycomputer name):~>. When I open the MyComputer Desktop Icon the drives:/ folder is empty. I've been to Novell looking for paid support but the prices start a $500. Does anyone know what's going on here? Does anyone know how to figure out what's going on here? Does anyone know how to fix this? This does not seem to be file system problem according to the rescue script on the installation dvd. So why is fsck failing at boot? Why does resierfsck --fix-fixable complain it can't fix the file system because it was mounted with write permissions? Is it supposed to be unmounted to fix? Why does mount show two instances of my root partition? How come umount doesn't get rid of one of them? Thanks, Jerome
On Thursday, 7 October 2004 04.30, Jerome Lyles wrote:
I was following Carl Spitzer's advice concerning getting my firewire drive to let me write to it(It worked,thanks Carl). He also noticed that in my fstab I had my floppy and dvdrecorder devices under subfs and he suggested I add 'users' to the fs= part of the media under subfs,which I did using vim. But I still couldn't write to the floppy even though I owned it. So I decided to run SuSEconfig-no help. Then I decided to reboot and this has led to a serious problem for me.
When I reboot I get the message that: ' fsck failed, my / partition is being mounted read-only' and I must login as root. However when I run mount I get this: (none):~#mount
hm, why does this say (none)? It should say (repair filesystem)
/dev/hdb2 on / type(reseiserfs) (rw) /dev/hdb2 on / type(reseiserfs) (rw) Two entries for the same file system! And the filesystem is mounted (rw).
Makes very little sense.
Running: (none):~#fsck fails and I'm told to run it manually. But when I run fsck.reseiserfs --fix-fixable I get this: (none):~#fsck.reseiserfs --fix-fixable Partition /dev/hdb2 is mounted with write permissions, cannot check it. When I mount the system read-only it gets checked but nothing gets fixed.
ps At the fsck failed login prompt I log in as root, enter init5 and restartx.
That is not a good idea. Few things will work, as you've noticed. You really can't afford to ignore a corrupt root file system. If fsck fails it is for a reason.
This does not seem to be file system problem according to the rescue script on the installation dvd. So why is fsck failing at boot? Why does resierfsck --fix-fixable complain it can't fix the file system because it was mounted with write permissions? Is it supposed to be unmounted to fix?
Yes, although it can work if it's mounted read only.
Why does mount show two instances of my root partition? How come umount doesn't get rid of one of them?
All excellent questions that I think will be hard to answer without seeing the system. My suggestion would be to boot the rescue system on the DVD and run reiserfsck on /dev/hdb2 without mounting it
On Wednesday 06 October 2004 07:10 pm, Anders Johansson wrote:
On Thursday, 7 October 2004 04.30, Jerome Lyles wrote: Then I decided to
reboot and this has led to a serious problem for me.
When I reboot I get the message that: ' fsck failed, my / partition is being mounted read-only' and I must login as root. However when I run mount I get this: (none):~#mount
hm, why does this say (none)? It should say (repair filesystem)
It still says (none).
/dev/hdb2 on / type(reseiserfs) (rw) /dev/hdb2 on / type(reseiserfs) (rw) Two entries for the same file system! And the filesystem is mounted (rw).
This is still true also. Is there some way to check to see if /dev/hdb2 is really mounted twice? Is there some way to check to see if mount is broken?
Makes very little sense.
ps At the fsck failed login prompt I log in as root, enter init5 and restartx.
That is not a good idea. Few things will work, as you've noticed. You really can't afford to ignore a corrupt root file system. If fsck fails it is for a reason.
Maybe it would help if someone can tell me how to tell fsck to tell me why its failing, some kind of verbose mode boot option for fsck .
This does not seem to be file system problem according to the rescue script on the installation dvd. So why is fsck failing at boot? Why does resierfsck --fix-fixable complain it can't fix the file system because it was mounted with write permissions? Is it supposed to be unmounted to fix?
Yes, although it can work if it's mounted read only.
Why does mount show two instances of my root partition? How come umount doesn't get rid of one of them?
All excellent questions that I think will be hard to answer without seeing the system.
My suggestion would be to boot the rescue system on the DVD and run reiserfsck on /dev/hdb2 without mounting it
Using the rescue system on unmounted partitions I've run: e2fsck -f -c -v -B /dev/hda1 (my MBR is on this drive), checks out clean. I've run:reiserfsck /dev/hdb2 (/), reiserfsck --fix-fixable /dev/hdb2, reiserfsck --rebuild-tree /dev/hdb2,reiserfsck --rebuild-db /dev/hdb2 and reiserfsck --fix-fixable /dev/hdb2 again. There were some problems, all were corrected. Both filesystems have a clean bill of health according to reiserfsck and e2fsck. I found a clue, maybe: /etc/init.d/rc*.d/: more than one link for service: syslog,nmb, ntop, snort, portmap, resmgr, smbfs, splach_early, nfslock, nfs, nfsboot, acpid, alsasound, fbset, postgresql, running-kernel, sshd, splash, atalk, hwscan, xntpd, cups, ypserv, nscd, smb, nfsserver, postfix, apache2, cron, splash_late; this is a partial list. Should I bother trying to side step this problem by copying my root partition to my external firewire drive? If yes, what dd or rsync command would do this for me? Someone mentioned dd_rescue but there are no info or man pages for it. Also, should I move the MBR from hda to hdb since my / is there? If yes, how do I do that? Or should I put a clean install on /dev/hda1 or my extermal firewire drive and rsync my /dev/hdb2 / partition with the clean install? Both reiserfsck and e2fsck in rescue mode say /dev/hdb2 and /dev/hda1 respectively are clean. If this is true, what other part of the system could be causing this behavior? Thanks, Jerome
participants (2)
-
Anders Johansson
-
Jerome Lyles