Hard disk failure

[sendhilk]

Hi,

I have a RS/6000-7013 Server. It gives a frequent error messages that

PROBLEM WAS DETECTED ON Wed Jun 20 03:01:04 2001 801014

The Service Request Number(s)/Error Code(s).
706-128: Error log analysis indicates a hardware failure.

Probable Cause or Causes:
- 100% hdisk2 00-08-00-2,0 4.5 GB 16 Bit SCSI Disk Drive
I see that the volume group to which this disk is attached in still active and all the file systems on this disc are still mounted.however to be safe i want to replace this disc.

This is how i plan to do it.
Attach a new disk to the existing system.add it to the volumegroup.
varyoff the volumegroup.
use migratepv to migrate the contents of hdisk2 to the new disk.
Is this procedure ok , does anyone have any better idea??
thanks,
sendhil

 

[crowe]

if it is just a data disk, I would get a tape backup of it ASAP. Failing drives are flakey at best. It could last 6 more months, or 6 more minutes. This way you are not relying on the drive "living" until you migrate the data from disk to disk.

 

[elvism]

What I recommend is to do a mksysb of the system and when you replace the hard disk, restore the system from the mksysb.

Cheers.

 

[singo]

I agree with last message. That is what we do. Just make sure your mksysb on tape is reliable and be sure to select all HDisks when restoring the image. On AIX 4.3.3 you can create a mksysb on CD now but to do it you have to get the CD write filesets off the net.

 

[MikeLacey]

I must be really lazy then - I always get the new disk installed as well - and then dd the data between the old and new disks before removing the old one.

(I *do* take a backup of the old disk as soon as I can though - as crowe says it would be embarassing if the old disk died completeley *just* before you xferred the data) Mike
michael.j.lacey@ntlworld.com
Email welcome if you're in a hurry or something -- but post in tek-tips as well please, and I will post my reply here as well.

 

[Skeater]

Mksysb's are very good if that physical volume you are having a problem with is part of the "rootvg" volume group.
Mksysb only backs up rootvg.

If it is not, you can do a volume group save through "smit".

Make sure you remove the description of the physical volume you are replacing BEFORE you put in a new one, or you will have problems removing it.

Once this is done you can do a restore volume group through smit.

 

[sendhilk]

Thanks a lot for all the input.
since i didnt have any slot left i could not add an extra disc and go a migratepv as i had planned.
So this is what i did.
I removed the defective disc from the volumegroup using reducevg.
Then i shutdown the machine, removed the defective disc and then added new disc.
booted the machine and then added the new disk to the volumegroup. recreated thelogical volumes and filesystems on the hard disk and restored all the data back to the filesystems.

Sendhil