Does a raw disk auto detect Dynamic LUN expansion

[John Jimenez]

I have a LUN on HP-UX 11.23 that is on a Oracle RAC file system, not LVM. I expanded and doubled the size of a LUN on EMC CLARiiON and created a METALUN, but the Oracle Admin is having issues recognizing the changes.

I was told that the raw disk auto detects the changes, but I have not found any documentation that supports this. I have read vgmodify, and it sounds like the vgmodify might change the size on the /dev/rdsk/cXtXdX.

Has anyone used dynamic LUN expansion on a Oracle RAC file system? Or does anyone know if I can check and see if the raw disk has changed?

[John Jimenez]

FYI I actually have both LVM and Oracle file systems on thie RP7420 running 11.23, but he one I need to increase is a Oracle file system LUN

[Tim Nelson]

run disk info on the device file.

diskinfo /dev/rdsk/cxtxdx

[Steven E. Protter]

Shalom,

In general lun expansion is not detected.

The best thing to do is present a new LUN of larger size and migrate the database.

SEP

[John Jimenez]

Tim,
Thank you. Yes that showed me that the Disks are still at 50 gig on both servers.

maxux1:[/dev/rdsk]
# diskinfo /dev/rdsk/c15t6d3
SCSI describe of /dev/rdsk/c15t6d3:
vendor: DGC
product id: CX500WDR10
type: direct access
size: 52428800 Kbytes
bytes per sector: 512


# diskinfo /dev/rdsk/c15t6d3
SCSI describe of /dev/rdsk/c15t6d3:
vendor: DGC
product id: CX500WDR10
type: direct access
size: 52428800 Kbytes
bytes per sector: 512

[John Jimenez]

Hi Steven, I am currently testing on a test LUN.
So there is really no way to have the raw disk recognize the changes?

A few months ago the Oracle admin asked for 500 gig to migrate data. I asked him how he wanted it spread out. He asked for two 200 gig and one 100 gig. one 200 gig and one 100 gig were merged. I guess Oracle balances data on these, but it looks like Oracle does not look at the disk sizes, only that it wants to equally spread out data evenly on both disks. We keep having issues ever 1.5 to 2 days. We were hoping to increase this from 100 gig to 200 gig to match the other disks. If we have to restore to migration, with my available space, I will have to shuffle things on the SAN itself to make it happen. This is doable, but I was hoping we were looking for a simpler solution and hoping we could take advantage of using METALUN.

[John Jimenez]

Hi Steven,
The migration you suggested, woud be done on the Oracle side, right? Or do you mean the EMC side? If it do it on the EMC side, I do not think the O.S. would still recognize it.