request for procedure for repalce MSA30 HDD

[Md. Khairul Alam Prodha_2]

Dear all,

We have rp4440, MSA30 (14 unit 72GB SCSI Hard Disk), OS HP-UX 11i v11.11 server running oracle 9i Database.

We want to replace 6 units 72GB Hard Disk (slot 1,2,3 & 7,8,9)with new 6 units 300GB Hard Disk and wants to software mirror.

We will take make_tape_recovery of OS and Cold Backup of Database before start the job.

Can any one provide us detail procedure for replace the 6 units 72GB HDD by new 300GB HDD?

Thanks
Khairul

[Steven E. Protter]

Shalom,

I sure hope you were not booting off the failed disk. The make_tape_recovery will only cover the disks on your rp4440 server, which probably does not boot off the MSA-30.

vgdisplay -v /dev/vg00

lvdisplay -v the logical volumes shown in the prior command.

Shut the database and do a vgexport on the volume group it is in before starting the procedure.

Note the replacement procedure iteslf is easy. The MSA-3000 is a hot swap disk holder and you pop out the old disk and insert the new one.



After that, your database will be gone and you'll need to pvcreate the new disk(s) probably with the -f parameter.

Then you vgimport (if you did the vgexport) or use lvcreate to build new logical volumes, newfs the filesystem and restore your cold database backup.

SEP

[Andrew Young_2]

Hi Steven

If the disks are configured as a JBOD (just a bunch of disks, no RAID controller) and are mirrored then I would recommend breaking the mirror with lvreduce -m 0 , moving the data of the first disk using pvmove, using vgreduce to remove the disk from the volume group, replacing the disk, pvcreate, vgextend and then pvmove the data onto that disk and then doing the remaining disk usuing the same mechanism.

If there is no mirror then you stand the risk of losing data unless you can free up at least one disk, otherwise its a case of backup and restore. make_tape_recovery will not do it for you, you will probably need to use tar or cpio.

If there is a RAID controller involved things will get messy since you are replacing dissimilar disk sizes. My suggestion would be to backup the data, replace all the disks, rebuild the RAID array and restore.

Regards

Andrew Y