TCS V5.1 quorum disk; can not create

From: Cable, Randall <Randall.Cable_at_COMPAQ.com>
Date: Wed, 22 Aug 2001 19:08:03 -0500

Hey now managers! I could use a little assistance, if you please. I am an
early adopter for the Tru64 V5.1a EFT, because we need the LAN interconnect.
I have loaded the OS and everything was coming along until I attempted to
run clu_create. This command fails when trying to configure the quorum disk.


> excerpt from clu_create
Creating required disk labels.
  Creating disk label on member disk : dsk0
  Initializing cnx partition on member disk : dsk0h
  Creating disk label on quorum disk : dsk9
disklabel: write boot code: I/O error

Error creating disk label on quorum disk : dsk9c
  Initializing cnx partition on quorum disk : dsk9h

Error: clu_create: Failed relabeling disks.

Error: clu_create: Failed to create a cluster.

No sweat, select no in clu_create and then do it manually later with
clu_quorum -d add dsk2 1. As I understand it, this should add /dev/disk/dsk2
as the quorum disk, utilizing only the h partition.

> Error, when clu_quorum run manually
# clu_quorum -d add dsk2 1
Collecting quorum data for Member(s): 1

disklabel: write boot code: No such device

*** Error ***
Cannot create disk label on quorum disk : dsk2c

*** Error ***
Cannot configure quorum disk.

I tried manually labeling the h partition as cnx and this works, but the
script then marks all partitions as unused before blowing chunks.

I think I have found a couple of errors, undefined variables in
/usr/sbin/clu_quorum, that cause the script to blow up the same way.

My build is from February, even though we just ordered and received the EFT.
I have been told that there have been three or four subsequent revisions to
this code, but none of them made it into the EFT, for various reasons $$$.
Is there any work-around for this problem. I am proceeding with my cluster
installation, realizing that without a quorum disk, the config will be
unstable.

The quorum disk I am trying to use is a RZ1CD-CS and I have tried two
different disks, in case that was an issue, with the same result. Could
possibly be a problem with the BA356, but not sure how to identify it. The
disks are seen by the system and I can read and zero the disklabel.
My environment is the following: Tru64 5.1a EFT/TruCluster 5.1 without
patches (yet) 1xDS20/128MB, 1xDS10/384MB, 1xBA356-KH Storage shelf,
2xKZPBA-CB with y-cable/termination, No MC; using LAN Interconnect, Cluster
root/usr/var
Any help greatly appreciated.

Thanks!
rrc

Randall Cable
Professional Services - North America.
COMPAQ Services
- Phone at Compaq: (719) 592-7114
- Fax: (719) 592-7159
mailto:Randall.Cable_at_compaq.com
Received on Thu Aug 23 2001 - 00:09:50 NZST

This archive was generated by hypermail 2.4.0 : Wed Nov 08 2023 - 11:53:42 NZDT