ITEM: AI1268L

machine hangs importing nonrootvg and locks up

Question:

Desc:   Customer's machine states 'failure importing medvg1' during
        normal boot.  The machine eventually boots to a login, but
        after logging in, every command gets killed due to lack of
        paging space.
                There are 2 disks in the system:

                hdisk0  rootvg  16MB paging space
                hdisk1  medvg1  100MB paging space

        The system has about 96MB of memory, so without the paging space
        from medvg1, the system fails to execute commands after login.

Act:    Booting in maintenance mode, 'lspv' lists the 2 disks and the
        volume group names correctly, but the volume group medvg1 is
        not varied on.  We ran:

                exportvg medvg1
                importvg -y medvg1 hdisk1
                varyonvg medvg1
                synclvodm -v medvg1

        which worked fine, but trying to mount the 1 filesystem in medvg1,
        /m2 fails with 'mount: invalid argument'.

        'fsck -y /dev/lv01' fails with a message 'unable to replay log:0'

        /etc/filesystems looks good and 'lqueryvg -Atp hdisk1' works fine,
        but it seems like the lv is damaged beyond repair.

        /etc/IPL_rootvg was missing so we recreated it with:

                        mknod IPL_rootvg c 10 0
                Then tried running logform and even re-creating the jfslog, but
        the filesystem will not pass an fsck.

                /usr/sbin/hdf /dev/hd4 1000 100
                /usr/sbin/hdf /dev/hd4 1F00 100

        show that the superblock and back-up superblock are corrupted.

        We removed the logical volume and tried to reboot in normal,
        but this failed the same way.

Next:   We booted in maintenance and exported the volume group medvg1.
        Then added a separate 48MB of paging space to rootvg and rebooted.
        We also rebuilt the ODM with:

                 ODMDIR=/dev/objrepos
                 mount /dev/hd4 /mnt
                 mount /usr

                 mkdir /mnt/etc/objrepos/backup         (copy current Cu info)
                 cp /mnt/etc/objrepos/Cu* /mnt/etc/objrepos/backup

                 cp /etc/objrepos/Cu* /mnt/etc/objrepos (copy over RAM ODM))

                 
                 cd /
                 /etc/umount all

                 bosboot -a -d /dev/hdisk1
                 shutdown -Fr

        This time the system booted to a login and commands work.  We
        created a new volume group on hdisk1, re-created the paging and
        the filesystem m2.  Next we restored the tty and printer info
        from the ODM with:

                export ODMDIR=/etc/objrepos/backup

                odmget -q "name like 'tty*'" CuAt > TTYs
                odmget -q "name like 'tty*'" CuDv >> TTYs
                odmget -q "value3 like 'tty*'" CuDvDr >> TTYs

                
                odmget -q "name like 'lp*'" CuAt > LPs
                odmget -q "name like 'lp*'" CuDv >> LPs
                odmget -q "value3 like 'lp*'" CuDvDr >> LPs

                export ODMDIR=/etc/objrepos

                odmadd LPs
                        (removed all entries for tty0 from TTYs since it had
                         already been automatically configured)
                odmadd TTYs

Next:   Steve will restore the data for /m2 and run bosboot

        closing call



Support Line: machine hangs importing nonrootvg and locks up ITEM: AI1268L

Dated: November 1995 Category: N/A

This HTML file was generated 99/06/24~13:30:27

Comments or suggestions?
Contact us