System panic after doing a dd to new disk - VFS_MOUNTROOTS Failed

[Norman Woo]

Folks

After doing a copy from a boot disk to a new disk:

dd if=/dev/dsk/c0t6d0 of=/dev/dsk/c0t5d0 bs=1024k conv=noerror

and swapping out the bad boot disk (c0t6d0) and replacing it with c0t5d0, I get a system panic with the message VFS_MOUNTROOTS failed: Need Drivers?

and then it goes and starts dumping memory to disk and starts the boot process over again.

I'm not able to get into single-user mode nor maintenance mode.

I put back the bad boot disk and it is able to boot OK but fails when fsck encountered a bad block, which I cannot bypass.

I thought doing a dd would create a clone of the boot disk. DD dd reported some read errors.

Thank

[T G Manikandan]

Though I have given a document but still the problem is that

dd checks for the bad blocks and probably the clone of the boot disk is not proper.

check for chapter three from the attachment.

[Ravi_8]

Hi,

dd can be used to clone the root disk, if the root disk is faulty then cloned disk will not be having data in that particular block.

since this is h/w problem, my recommendation is to cold install the OS on the new disk and extract the data from backup

[Ravi_8]

Hi,

dd can be used to clone the root disk, if the root disk is faulty then cloned disk will not be having data in that particular block.

since this is h/w problem, my recommendation is to cold install the OS on the new disk and extract the data from backup

if you have ignite baclup, it could help

[Ravi_8]

Hi,

dd can be used to clone the root disk, if the root disk is faulty then cloned disk will not be having data in that particular block.

since this is h/w problem, my recommendation is to cold install the OS on the new disk and extract the data from backup

if you have ignite backup, it could help

[Ravi_8]

Hi,

dd can be used to clone the root disk, if the root disk is faulty then cloned disk will not be having data in that particular block.

since this is h/w problem, my recommendation is to cold install the OS on the new disk and extract the data from backup

if you have ignite backup, it could help

[Ravi_8]

Hi,

dd can be used to clone the root disk, if the root disk is faulty then cloned disk will not be having data in that particular block.

since this is h/w problem, my recommendation is to cold install the OS on the new disk and extract the data from backup

if you have ignite backup, it could help

[Ravi_8]

Hi,

dd can be used to clone the root disk, if the root disk is faulty then cloned disk will not be having data in that particular block.

since this is h/w problem, my recommendation is to cold install the OS on the new disk and extract the data from backup

if you have ignite backup, it could help

[Norman Woo]

Hi

Wow, thanks for the detailed answer.

One thing though, if my "bad" boot disk which only had 1 bad block under /usr/lib was still able to boot correctly, shouldn't the new disk also boot correctly? I was planning to dd the old to new and then restore the /usr directory.

Another thing is that the current boot disk (c0t6d0) contains logical volumes lvol1 through lvol8. It is lvol7 that contains the bad block which happens to contain the /usr stuff.

I've fsck'd the other volumes and they are OK. I am able to mount them to their respective mount directories (/var, /tmp, /opt and so on). Only /usr cannot be mounted as fsck fails with a bad block read. I was able to force a mount of lvol7 to /usr and then dump out the data to tape. There were some problems with the /usr/lib/lib* libraries and those I can recover from another 10.20 system.

My alternative solution is to newfs the lvol7 (thereby marking the bad block) and then restore the /usr data from disk.

System is HPUX 10.20 running on HP9000/735. All disks are in HFS format