Sep 13 12:18:56 sfe fsck: /dev/VolGroup00/LogVol00 is mounted. Sep 13 12:18:56 sfe fsck: e2fsck: Cannot continue, aborting. Sep 13 12:18:56 sfe rc.sysinit: Checking root filesystem failed
I had updated the box a few days ago but I had not rebooted with the new kernel. After the reboot, I get the root filesystem is not marked clean message and also that it was not mounted read-only. I supplied the root password, remounted read-only, did the obligatory fsck and hit ctrl-D.
Guess what? I get the same message! In addition, there were no error messages during the fsck to boot.
Anyone else encounter this?
On Wed, 2006-09-13 at 14:30 +0800, Feizhou wrote:
Sep 13 12:18:56 sfe fsck: /dev/VolGroup00/LogVol00 is mounted. Sep 13 12:18:56 sfe fsck: e2fsck: Cannot continue, aborting. Sep 13 12:18:56 sfe rc.sysinit: Checking root filesystem failed
I had updated the box a few days ago but I had not rebooted with the new kernel. After the reboot, I get the root filesystem is not marked clean message and also that it was not mounted read-only. I supplied the root password, remounted read-only, did the obligatory fsck and hit ctrl-D.
Guess what? I get the same message! In addition, there were no error messages during the fsck to boot.
Anyone else encounter this?
I wish I could offer more. I reply only because I know what it is like when no one responds. I have not had any problems similar to what you describe. If it would be helpful, I'll be glad to compare/discuss any differences between our LVM and/or boot setups to see if we can ID the problem.
Have you done an ls -a on that file system? I seem to recall seeing some hidden files that can lay around that are related to fsck auto-check on boot? Not sure about that.
<snip sig stuff>
-- Bill
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Feizhou wrote:
Guess what? I get the same message! In addition, there were no error messages during the fsck to boot.
While I've never had this happen to me, I'd try two things:
1) Ensure that no advisory file exists to cause it. e.g. /forcefsck *should not* exist.
2) Boot on a CentOS install CD with a "linux rescue" command, then run the fsck on the offending partition. That may clean things up for you.
Barry
Barry L. Kline wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Feizhou wrote:
Guess what? I get the same message! In addition, there were no error messages during the fsck to boot.
While I've never had this happen to me, I'd try two things:
- Ensure that no advisory file exists to cause it. e.g. /forcefsck
*should not* exist.
- Boot on a CentOS install CD with a "linux rescue" command, then run
the fsck on the offending partition. That may clean things up for you.
I have looked for all the files checked by rc.sysinit...
All I can find is an empty /halt...probably left because I hit 'init 3' after I got fed up with the endless loop.
I have finally found where the problem is apparently...
I had these options for / in fstab:
rw,acl
This meant that / was mounted rw,acl by the time rs.sysinit was ready to call fsck -T -a on it and since it was not mounted read-only or since it was mounted (that's the impression I get from the error message) the script goes into repair mode for root or just reboots.
So this really has nothing to do with LVM :P
Has anyone tried to get acl on root?
-
Barry L. Kline -
Feizhou -
William L. Maltby