disk is currupted

[sachit patil]

HI all ,
I have one root disk and it got currupted and i installed new disk and i don't have backup of previous disk .can i get all missing vgs on this disk

[A. Clay Stephenson]

Let me see if I understand this. You have an unmirrored root (boot) disk for which you have no backups? You want to know if you can get all the missing vg's on this (new?) disk. Well, there can be at most one VG on a given disk. In any event, if you didn't bother to mess with those silly, unneeded backups then the answer for you is NO. Moreover, conventional backups for the root disk are not sufficient because you must first install on OS before you can run tar, cpio, or frecover. What you should have done is created an Ignite image which could have restored you without first having to load an OS. What you should have also done is have the disk(s) mirrored so that you wouldn't even have to shutdown to replace a failed disk. The technical term for your condition is "toast".

[Marvin Strong]

no backups = no restore
After you reinstall and reconfigure your system. It might be a good idea to make an ignite image or some other backup of the system this time.

[spex]

Clay -

Are you sure you the system would stay up after having the primary root disk fail?

I just added a mirrored root disk to our rp4440. A member of HP's disk storage team told me that in the event of the primary volume crashing, the server would most likely go down, but upon reboot, the system would boot off the mirrored drive (as an alternate boot path).

Maybe this behavior is different for different servers.

PCS

[Patrick Wallek]

That's the whole point of mirroring. If the primary fails, the system keeps running on the secondary. I have had root disks fail and the system remain up and running. You then replace the disk, assuming you have hot-swap disks, while the system is running and no one need ever know you had a disk failure.

[Bill Hassell]

And to amplify what Patrick said: the secondary disk mirror will take over automatically when the primary failure is detected. Now there can be some corner cases:

- if the two disks are on the same bus and the failed primary trashes the bus so neither disk will work.

- the failure on the primary disk is subtle (not a simple head crash) and hangs in an unexpected way.

Of course, any mirror disk design for production will ALWAYS use separate I/O cards for each disk to avoid the problem.

[TwoProc]

Addendum to Bill's note on separate I/O cards:

You must watch out for this yourself, as HP has happily shipped me multiple servers (including SD and RP) with a single I/O card (with two separate scsi busses on it) for both mirrors of root, even though they put two separate cards in the I/O slots (for this purpose as per the order). So, if your machine came pre-loaded with HPUX, it may not be configured correctly, and it's up to the sysadmin to make sure that it is.

[A. Clay Stephenson]

Yes, I am quite sure that the system will stay up. I literally cannot remember the last time I shutdown to replace a disk (boot disk or otherwise) but I'm sure that it has been at least 7-8 years. Perhaps "The Force" was with me but I have never had a hang of any kind as a result of a failed disk. Regardless of how the box was ordered and configured, I always assumed it was messed up so my first action on any new box is to load the OS the right way -- meaning, of course, the way I would load it. Sometimes, I turn this task over to another admin so that he can do it "correctly".

[Bill Hassell]

Back to the original problem...

If the boot disk is corrupted and will not bootup, you might contact www.ontrack.com to see if they can recover your data. It is very expensive (thousands of US dollars) but they have very specialized equipment to read disks with bad spots or misaligned heads. They cannot, however, recover data which was erased or replaced by writing all over the disk.

[sachit patil]

hi bill
I installed hp os and reconfigure it agin and solve that problem
thanks
regards
sachit patil