Hypervisor Dump Retrieval Procedure; Hypervisor Dump Indications - IBM eserver p Series Service Manual

Hypervisor Dump Retrieval Procedure

This procedure defines the requirements and methods for retrieving a hypervisor (LPAR) dump file
following a system crash. This procedure is only applicable if the managed system had been running as a
partitioned system before it crashed.
This procedure requires the managed system to be rebooted as a partitioned system. Retrieval must then
take place from an AIX partition.

Hypervisor Dump Indications

If the system op-panel is displaying error code 20D00997, 20D00998, or 20D00999, this is an indication
that a system crash within the runtime firmware has occurred.
If error code 20D00997, 20D00998, or 20D00999 is in the service processor error log, this is also an
indication that a system crash within the runtime firmware has occurred. The following is an example of a
detailed service processor error log entry with error code 20D00999:
05/29/2001 21:06:12
Error code: 20D00999
Detail: A2C1
SRC
--------------------------------------------------------------
word11: B1194690
word14: 00000000
word17: 00000000
20D00999
After a system crash has occurred, a check should be made to see if a hypervisor dump has been
collected. This is described in the following steps.
Retrieving the Hypervisor Dump file
Once it has been determined that the managed system has crashed and a hypervisor dump exists, or the
user wishes the test for an existing dump file, the managed system must be rebooted as a partitioned
system. A partition or partitions running AIX must then be activated. Once rebooted, any partition running
AIX may be used to retrieve the dump file.
1. Install the devices.chrp_lpar.base.ras Fileset.
In order to perform the following task, the devices.chrp_lpar.base.ras AIX fileset must be installed and
committed. To determine if the package has been previously installed, enter the following command on
the AIX command line:
lslpp -l | grep devices.chrp_lpar.base.ras
If the output on the AIX console is something similar to:
devices.chrp_lpar.base.ras
the devices.chrp_lpar.base.ras fileset is already installed and no further action for step 1 is necessary.
If no output is returned, the devices.chrp_lpar.base LPP will have to be installed from the AIX
installation media. This must be done by the customer or system administrator.
Once the fileset is installed, the dump file is ready to be retrieved.
2. Enter fetchdbg -k at the command prompt.
This should produce the following output:
Retrieving Dump File ............Dump read successful
Operating System Terminated with Error String
word12: 04A0005D word13: A2C1C0FA
word15: 00007701 word16: 00000008
word18: 00000000 word19: 00000000
5.X.X.X COMMITTED CHRP LPAR RAS Support
Chapter 5. Error Code to FRU Index
465