Red Hat Bugzilla – Bug 10804
Problems with fsck on booting
Last modified: 2016-10-18 12:35:39 EDT
After setting up 6.1 I was continualy getting put into maint mode. I'd
fsck the disks reboot which about 50% of the time on exit got through the
boot. After a few days when booting I noticed on the terminal that booting
messages showed the -C option with -C0 which was odd. I check the man page
and there was not -C0 option. I checked rc.d rc.sysinit and there was only
-C no -C0. I deleted the -C and have not had a problem boot since.
Somehow a 0 got inserted into the -C to form -C0.
I've been able to repeat the problem by inserting or deleteing the -C option in
rc.sysinit. When not present I boot clean. When -C is inserted most of the
time I get put in maintenance mode. With -V added show's the fsck command on
the terminal as having -C0.
my fstab is:
/dev/hda2 / ext2 defaults 1 1
/dev/hda1 /boot ext2 defaults 1 2
/dev/cdrom /mnt/cdrom iso9660 noauto,owner,ro 0 0
/dev/hda6 /opt ext2 defaults 1 2
/dev/hda7 /usr ext2 defaults 1 2
/dev/hda5 swap swap defaults 0 0
/dev/hdb1 /r ext2 defaults 1 2
/dev/hdb2 /spare ext2 defaults 1 2
/dev/hdc1 /s ext2 defaults 1 2
/dev/hdc2 /t ext2 defaults 1 2
/dev/fd0 /mnt/floppy ext2 noauto,owner 0 0
/proc proc defaults 0 0
/dev/pts devpts gid=5,mode=620 0 0
There is a bug in the e2fsprogs that shipped with 6.1; you
need to grab the errata version.
The error is still happening after the errata update.
The error reported on boot is 10 which I think is a
usage error. The error reports that it executed
fsck.ext2 but I don't think that is the case.
For example if I umount /dev/hdb2 and fsck /dev/hdb2
that works. If I fsck -C /dev/hdb2 that works.
If I fsck -C0 /dev/hdb2 I get a usage error (10).
If I fsck.ext2 -C0 /dev/hdb2 it works.
It's acting like fsck and fsck.ext2 are getting randomly
invoked on boot and if fsck gets invoked I get the
error and if fsck.ext2 gets invoked its ok.
The man page tells me fsck is the front end for
fsck.fstype so fsck must do some evaluation of the disks
and on that basis calls fsck.fstype from fsck
with the correct .fstype.
It seems that is not happening consistantly. The way
its acting is it randomly runs fsck and at other times
If I remove the -C option in rc.sysinit I boot
without getting put into maintenance mode.
Note: the error report said that it ran fsck.ext2 but
its not acting that way. It acting like it called
fsck and no .fstype got evaluated or attacked to the
Please disregard the last update. USER ERROR
on picking up the correct update. Sorry.
Commit pushed to master at https://github.com/openshift/origin
Merge pull request #11222 from jim-minter/issue10804
Merged by openshift-bot