Problem with interactive fsck at bootup

holst · 02-26-2005, 12:36 PM

Problem with interactive fsck at bootup. To begin with here is my fstab:

Code:

  LABEL=BOOT      /boot           ext2    noauto,noatime  0 0 
  LABEL=ROOT      /               auto    noatime,errors=remount-ro 0 0 
  /dev/hda2               none            swap            sw      0 0
  /dev/cdroms/cdrom0      /mnt/cdrom      iso9660         noauto,ro       0 0

If the ROOT disk is not clean - and need a fsck - the bootup sequence will abort and start asking interactive questions to the user. This is no good as i CAN'T ANSWER THEM. I only have access to the machine via remote shell (SSH). Hence I need to machine to continue to try and boot into network mode so i CAN access the machine and run fsck on the disk (when the disk is in read-only mode). I also would be nice if the system could give me some indication that the disk is not clean and need a fsck when I login by remote.

jailbait · 02-26-2005, 02:07 PM

"the bootup sequence will abort and start asking interactive questions to the user. This is no good as i CAN'T ANSWER THEM."

fsck has an automatic option which turns off the questions. See man fsck. You can find the boot scripts which run fsck and add the -a option to the fsck command. That way boot can procede without your intervention. In SuSE the following files would need to be changed:

/etc/init.d/S04boot.lvm
/etc/init.d/S05boot.localfs
/etc/init.d/S06boot.crypto

I don't know what the corresponding Gentoo files are called.

----------------------------
Steve Stites

holst · 02-26-2005, 02:46 PM

Its already full of -a

Code:

sh-2.05b$ grep 'fsck' /etc/init.d/checkroot 
                if [ -f /forcefsck ]
                  ewarn "A full fsck has been forced"
                  fsck -C -a -f /
                  # /forcefsck isn't deleted because checkfs needs it.
                  fsck -C -T -a /
sh-2.05b$

Still no go. It stalls and asks for my confirmation to fix errors.

holst · 02-26-2005, 04:53 PM

I tried to replace "-a" with "-p" in the script and it SEEMS it works better now. I tried with this:

1. create a "/forcecheck" file to make sure the checkroot script was exected.
2. reboot.
3. ...wait for 10 minutes for disk to be scanned... (??? something took some time here atleast)
4. machine surfaced!

Entire script:

Code:

#!/sbin/runscript
# Copyright 1999-2004 Gentoo Technologies, Inc.
# Distributed under the terms of the GNU General Public License v2
# $Header: /home/cvsroot/gentoo-src/rc-scripts/init.d/checkroot,v 1.39 2004/04/21 17:09:18 vapier Exp $

depend() {
    before *
}

start() {
    local retval=0

    if [ ! -f /fastboot -a -z "${CDBOOT}" ]
    then
        ebegin "Remounting root filesystem read-only (if necessary)"
        mount / -n -o remount,ro &>/dev/null
        eend $?

        ebegin "Checking root filesystem"
        if [ -f /forcefsck ]
        then
            ewarn "A full fsck has been forced"
            fsck -C -p -f /
            # /forcefsck isn't deleted because checkfs needs it.
            # it'll be deleted in that script.
            retval=$?
        else
            fsck -C -T -p /
            retval=$?
        fi
        if [ "${retval}" -eq 0 ]
        then
            eend 0
        elif [ "${retval}" -eq 1 ]
        then
            ewend 1 "Filesystem repaired"
        elif [ "${retval}" -eq 2 -o "${retval}" -eq 3 ]
        then
            ewend 1 "Filesystem repaired, but reboot needed!"
            echo -ne "\a"; sleep 1; echo -ne "\a"; sleep 1
            echo -ne "\a"; sleep 1; echo -ne "\a"; sleep 1
            ewarn "Rebooting in 10 seconds ..."
            sleep 10
            einfo "Rebooting"
            /sbin/reboot -f
        else
            eend 2 "Filesystem couldn't be fixed :("
            /sbin/sulogin ${CONSOLE}
            einfo "Unmounting filesystems"
            /bin/mount -a -o remount,ro &>/dev/null
            einfo "Rebooting"
            /sbin/reboot -f
        fi
    fi

    # Should we mount root rw ?
    if mount -vf -o remount / 2> /dev/null | \
            awk '{ if ($6 ~ /rw/) exit 0; else exit 1; }'
    then
        ebegin "Remounting root filesystem read/write"
        mount / -n -o remount,rw &>/dev/null
        if [ "$?" -ne 0 ]
        then
            eend 2 "Root filesystem could not be mounted read/write :("
            /sbin/sulogin ${CONSOLE}
        else
            eend 0
        fi
    fi

    if [ "${BOOT}" = "yes" ]
    then
        local x=
        local y=

        #
        # Create /etc/mtab
        #

        # Clear the existing mtab
        > /etc/mtab

        # Add the entry for / to mtab
        mount -f /

        # Don't list root more than once
        awk '$2 != "/" {print}' /proc/mounts >> /etc/mtab

        # Now make sure /etc/mtab have additional info (gid, etc) in there
        for x in $(awk '{ print $2 }' /proc/mounts | uniq)
        do
            for y in $(awk '{ print $2 }' /etc/fstab)
            do
                if [ "${x}" = "${y}" ]
                then
                    mount -f -o remount $x
                    continue
                fi
            done
        done

        # Remove stale backups
        rm -f /etc/mtab~ /etc/mtab~~
    fi
}


# vim:ts=4

holst · 03-30-2005, 12:12 AM

nope. didnt work. same error one again! *goes jumping jackson in joy*