LUN size mismatch

[arkie]

Hi All,

In an EVA 4000, a Vdisk of size 2000 GB has been presented to a particular host. But, at the host end, the displayed size is not at par.

# diskinfo /dev/rdsk/c8t0d1
SCSI describe of /dev/rdsk/c8t0d1:
vendor: HP
product id: HSV200
type: direct access
size: 4244635648 Kbytes
bytes per sector: 512

This works out to approx 3.953125 TB (but desirably, should have been 1.953125 TB).

Why is there this size difference?

[Torsten. · Accepted solution]

Strange. Absolutely sure it is the same vdisk?

[arkie]

Yes Torsten,

There is only presented LUN.

[TTr]

I will ask the same question as Torsten. Are you sure you have the right LUN?

It maybe the only LUN in the server but was the right LUN presented from the array? Or are you sure it was created at the array as a 2000GB LUN? Ask the SAN team to verify again.

Do you have the EVA software install on the host? Check the LUN with evainfo.

[Torsten.]

AFAIR for this array 2TB is the maximum LUN size anyway. Strange.

"evainfo" would be interesting.


What is the host OS?

I remember some diskinfo patches years ago.

[singh sanjeev]

if you have doubt on diskinfo output ..check in SAM about the size

[Aungshuman Paul]

What is OS version and Operating environment ( Base, Mission Critical etc. )

[arkie]

I would like to come up with all the information at the earliest, but sorry! I have been in another customer site for the last few days.

By the way OS is HP-UX 11iv1 and OE is Base

Will await your responses as always....

Auf Wiedersehen!

[S. Ney]

Do you have access to HP's business support center? I'd get the evainfo tool for 11v1

http://h20000.www2.hp.com/bizsupport/TechSupport/SoftwareDescription.jsp?lang=en&cc=US&swItem=co-52747-1

If you can download run the evainfo command with -a or -d option.

Also not sure this thread would help you or not:
http://forums11.itrc.hp.com/service/forums/questionanswer.do?threadId=1307414

Lastly does your syslog or /etc/rc.log show any errors?

[TTr]

> Torsten: I remember some diskinfo patches years ago.

Indeed. From the PHCO_27168 description:

"When diskinfo(1M) is run on a large capacity disk, wrong value is printed for fields "blocks per disk" and "size".

No reboot required, the patch replaces the diskinfo binary.