Members may not be specified for preexisting volgroup

[Robert Hayden]

This is my first post on this particular listserv, so please bare with
me.  I have also read through the last few years of posts without
finding answers to my question.  So, I thought I would ask the
community before logging a support ticket with Red Hat.

I am trying to isolate a kickstart issue with RHEL 6.4 using anaconda
13.21.195.  Hardware is HP ProLiant BL460c G7. I have SAN storage and
an internal hard drive.  In the %pre script, I am removing the lpfc
kernel module and loading the hpsa module.  Only the internal drive is
available when I review the anaconda session in tty2.

Issue:  My kickstart session is aborting with the following message:

-------------start cut------------------
The following error was found while parsing the kickstart configuration file:

The following problem occurred on line 38 of the kickstart file:

Members may not be specified for preexisting volgroup

install exited abnormally [1/1]
The system will be rebooted when you press Ctrl-C or Ctrl-Alt-Delete.
----------------end cut---------------

My partition section in the kickstart is as follows with line 38 being
the volgroup line:

--------------start cut--------------------
bootloader --location=mbr --driveorder=sda --append="crashkernel=auto
rhgb quiet"
clearpart --none
part /boot --fstype=ext4 --onpart=sda1
part pv.008002 --onpart=sda2 --noformat
volgroup vg00 --pesize=32768 --useexisting --noformat pv.008002
logvol / --fstype=ext4 --name=rootlv00 --vgname=vg00 --useexisting
logvol /tmp --fstype=ext4 --name=tmplv00 --vgname=vg00 --useexisting
logvol /home --fstype=ext3 --name=homelv00 --vgname=vg00 --useexisting
--noformat
logvol /u02 --fstype=ext3 --name=diaglv00 --vgname=vg00 --useexisting --noformat
logvol /opt --fstype=ext3 --name=optlv00 --vgname=vg00 --useexisting --noformat
logvol /opt/stage --fstype=ext3 --name=optstagelv00 --vgname=vg00
--useexisting --noformat
logvol /usr --fstype=ext4 --name=usrlv00 --vgname=vg00 --useexisting
logvol /u01 --fstype=ext3 --name=u01lv00 --vgname=vg00 --useexisting --noformat
logvol /usr/local --fstype=ext3 --name=usrlocallv00 --vgname=vg00
--useexisting --noformat
logvol /var --fstype=ext4 --name=varlv00 --vgname=vg00 --useexisting
logvol swap --name=swaplv00 --vgname=vg00 --useexisting
-------------------------end cut----------------------

What I am attempting to do is install over a RHEL 5 machine,
reusing/formating core file systems like /etc and /var while keeping
user data file systems like /u01 and /opt intact.  The idea is to do a
variation of a fresh install instead of an in-place upgrade.  The
"upgradeany" option from RHEL 5 to RHEL 6 leaves too much junk
compared to the 4 to 5 jump.

The GUI version of anaconda will allow me to pick and choose which
file systems to reformat and which to skip.  That works fine for my
use case.  The attempt on the partitioning section  above was using
the anaconda-ks.cfg from the interactive installation as a starting
point.

I have completed a wide variation of combinations with the parameters
--noformat and --useexisting and have not hit the correct combination
yet.  I have also scanned the source code for anaconda and cannot find
the phrase "Members may not", so I am not even sure where the code is
aborting.

When I switch to tty2 for troubleshooting, the last file touched in
/tmp is the storage.log with multipath being used to scan devices.  I
see /dev/sda*, but /dev/mapper directory only contains "control"
special file.  Unsure if that is makes a difference, since I am using
the --onpart=sda2 for the volume group.

I am looking for suggestions on how to format the kickstart file to be
able to skip file systems on a fresh installation.  Or, some direction
in how to step through or debug anaconda to determine where in the
process it aborts.  This is an R&D machine, so I can pretty much toast
it on a whim.

Sincerely
Robert