ITEM: AP5124L

7137 installation corrupted the Pd* odm



Env:  He is at AIX 4.1.3 on a C20 attached to a 7137-413. He is using an
Enhanced SCSI-2 DE F/W adapter (FC 2412). He has 5-2.0GB disk drives in
the 7137 and is using RAID 5.

Question:  7137 disk will not become available

Answer:  In this case the 7137 installation currupted all Pd* odm entries
    below the 7137 entries.  See 2) for how to resolve this.

______________________________________________________________________

                Preliminary Problem Determination Steps
SE was getting an
   hdisk2 Defined      IBM 7137-413 RAID
   hdisk3 Available    Other SCSI Disk

Act: First checked his configuration, he has a cable with part number
46G4219 which is the correct cable for the 7137, the correct terminator,
and the correct diskettes for installation. Diskettes have the following
part numbers:

      PN\# 31H6842   for the Installation Diskette
      PN\# 31H6845   for the Microcode Diskette

He already had installed the Software for the 7137-413 and he said that
the Microcode on the 7137 was at a higher level than the microcde on the
diskette so he did not install the microcode on the diskette.

***************************Removing odm entries***********************
Tried to remove both 7137 hdisk definitions with the 'rmdev -dl \'
hdisk3 removed fine but  hdisk2 came back with a 
     0514-066 Cause not known 

After everything for hdisk2 was removed from the ODM, checked with the
'lsdev -Cc disk' command only showed his internal disks. I removed all
his 7137 software to try an reinstall it or start from scratch. Did
an 'rm -r 7137' in the /usr/lpp directory to remove the 7137 software.
Did the following steps to install the software:

      o mkdir 7137 - in the /usr/lpp directory
       o tar -xvf /dev/fd0 - to download the software from the Installation
                             diskette
         o ran the /usr/lpp/7137/arrayinst program
*************************************************************************

The arrayinst program completed successfully however during the
installation he received the following error:

      0518-507 odmget Cannot retrieve the object from PdAt. 5803 error

An 'lsdev -Cc disk' returned the correct configuration:
     hdisk2 Available     IBM 7137-413 RAID

And no hdisk3. However he IPLed the system and when it came back up he
had the same configuration as in the beginning. The 'lsdev -Cc disk'
showed the following:

   hdisk2 Defined      IBM 7137-413 RAID
   hdisk3 Available    Other SCSI Disk

Tried looking through his PdAt. Did an 'odmget PdAt|grep 7137' and this
returns the same error:

     0518-507 odmget Cannot retrieve the object from PdAt. 5803 error

Did an 'odmget PdAt|more' and let it run and it runs for a while an gets
the same error towards the bottom.

Advised that this appears to be a problem with his ODM. 

______________________________________________________________________

1)                  Determining if this is an ODM Problem

Act:  Since the customer has a 7137-413, he should be able to obtain
the information from the pre-defined database (PdAt) of the odm with
the following command:

 odmget -q deflt=7137413 PdAt

Here is what the output of that command should return:

PdAt:
        uniquetype = "disk/scsi/7137413"
        attribute = "model_name"
        deflt = "7137413"
        values = ""
        width = ""
        type = "R"
        generic = ""
        rep = "s"
        nls_index = 0

The above command core dumped for the customer.  Therefore the pre-
defined odm database has become corrupted.  At least for the 7137
portion.

I had the customer issue the following command and search for 7137:

odmget PdAt | more

The command failed with the following error message which indicates
that their is a problem with the PdAt:

0518-507 odmget Cannot retrieve the object from PdAt.

Normally, you would get the message "Pattern not found".

Then I had the customer issue the same command and search for bradley.
This time, he got the message "Pattern not found" and the same error
message as above.

This would lead me to conclude that the 7137 entries in the pre-
defined odm database must have gotten corrupted.

________________________________________________________________________

2)                   How to correct the ODM

NextAct:  Use the following commands to delete everything related to
the 7137 from the pre-defined odm database.

 odmdelete -o PdAt -q deflt=7137XXX
 odmdelete -o PdDv -q type=7137XXX

Where 7137XXX should be substitued by each of the 7137 models:
    7137412
    7137413
    7137414
    7137512
    7137513
    7137514
    7137090

Then follow the same procedures as defined in the previous response
for removing the 7137 software and then reinstalling it.  See 
Preliminary Problem Determination Steps.

_____________________________________________________________________

3)                            Further Check

Act:  Lee M. and I went through a process of elimination to see if we
could restore the Pd* files from /usr/lib/objrepos.  We were unable to
do so.  Thus, the customer will have to reinstall the system from a
mksysb backup.

Act:  The customer completed the reinstallation and had re-installed
the 7137 software and thought that the cfgmgr was taking too long but
the cfgmgr completed while I had him on the phone.  Also, I had him
look at the pre-defined odm database (PdAt) and it looked ok.  Then, I
had hime issue the command "lsdev -Cc disk" and the 7137 disk showed
up as Available.

NextAct:  Closing item per customer agreement.



Support Line: 7137 installation corrupted the Pd* odm ITEM: AP5124L
Dated: March 1996 Category: N/A
This HTML file was generated 99/06/24~13:30:26
Comments or suggestions? Contact us