SUMMARY: KZPCC drives not accessible with 5.1B/PK4 & TruCluster

• Previous message: Ron Bramblett: "Summary: 2nd Failed SCSI Drive in 2 weeks"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Date: Thu, 19 May 2005 10:41:39 -0400
To: tru64-unix-managers@ornl.gov

Thanks to everyone that replied to me.

I installed the patch in a non-roll (since the cluster is being rebuilt atm
anyway.) When I ran it, although I was presented with options to install for
Tru UNIX and Tru Cluster, the correct choice was only Tru Cluster, after
that the patch install worked fine.

Here are the responses I received:

--- Dan Price ---

Sheldon,

  Check out this ERP (Early Release Patch) page - search for KZPCC .

http://h30097.www3.hp.com/unix/EarlyReleasePatch-download2004.html

Dan

I got bit by this too. According to what I was told by HP, they added some
new functions in PK4 to the SCSI subsystem that the cluster components now
require. Unfortunately, they forgot to add those functions to the drivers
that handle local RAID controllers like the KZPAC's I had trouble with.

At the time I had to back down to PK3 for the project I was working on, but
I'm told that there is now a patch for this.

                                    - Bluejay Adametz

Sjolshagen, Thomas

Known problem (been known for months). Is a DRD/CAM interaction thing.

Check the Tru64 UNIX patch section of itrc.hp.com and install (*all* of)

the required ERP patches.

// Thomas

Polnick, Michael

There is a patch for this:

TCRKIT0023955-V51BB25-E-20041017.tar:

A dupatch-based, Early Release Patch kit exists for HP TruCluster Server
V5.1B-2
that contains solutions to the following problem(s):

Storage behind adapters such as KZPAC and KZPCC become inaccessible on
installing V5.1B PK4/BL25. The problem occurs because a new interface was
introduced in BL25 which requires the underlying driver to implement the
DEV_GET_EEI data structure. Some drivers, such as i2o and re_driver, have
not implemented this data structure, and consequently DRD incorrectly
interprets the response from these drivers.

Regards

Michael

Note: this patch has been superceded by a new patch.

- Zoong PHAM

IIRC, there is an issue with PK4 and the KZPCC. There is an ERP post PK4 for
it.

Plus there are 1 or more ERP post PK4 for Oracle as well.

And there are at least 2 ERP post PK4 for TruCluster.

I can get more details about it for you tomorrow.

Thanks so much for everyone's help!

Cheers,

Sheldon Lee-Wen
Unix Systems Engineer
Ottawa Data Center, Unix Services
CGI Information Systems and Management

• application/x-pkcs7-signature attachment: smime.p7s

• Previous message: Ron Bramblett: "Summary: 2nd Failed SCSI Drive in 2 weeks"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]