E10K hpost failure in final config.....

phase final_config: Final configuration...
Configuring in 3F, FOM = 114688.00: 8 procs, 14 Scards, 8192 MBytes.
FAIL Proc 0.0: Test FC I/O Init id=0x1F7 Bogus server_retval 83886080.
{0.2}   SysIO RAM Error, SBuffer_RAM_Diag,
{0.2}   address  00000000.000052b0
{0.2}   expected 00000000.00000000
{0.2}   observed 9016000a.dada02a0
{0.2}   xor      9016000a.dada02a0
{0.2} *** Error in SYSIO 0xd RAM init
Only 7 procs left, started configuration with 8!
Only 7 procs left, started configuration with 8!
Failures during final configure require reevaluation of best configuration.
Configuring in 3F, FOM = 100352.00: 7 procs, 14 Scards, 8192 MBytes.

If you are experiencing a failure that resembles the above, you may have a problem caused by a corrupt centreplane configuration. A boot of the domain will often result in a panic.

Escalations 525315 and 525459 refer.

DO NOT SWAP PARTS.

Engineers in EMEA should call the EMEA E10K Control Room.

The condition can be cleared by power cycling the centreplane (after shutting down all domains) and reconfiguring with hpost -C

The problem has been observed on platforms when hardware service or upgrades are taking place and system boards have been inserted or power cycled.

Note that the escalation states that the root causes are issues fixed by various SSP patches, implying that the best way to avoid the problem is to proactively ensure that SSP patches are at the latest level. Note also that installing patches while the platform is displaying the problem will not fix the problem. A centreplane reconfiguration is required.