After ILOM Firmware Update Error

Asset ID:	1-72-1525020.1
Update Date:	2017-10-05
Keywords:

Solution Type Problem Resolution Sure

Solution 1525020.1 : After ILOM Firmware Update Error

Applies to:

Sun Netra T5440 Server - Version All Versions to All Versions [Release All Releases]
Sun SPARC Enterprise T5240 Server - Version All Versions to All Versions [Release All Releases]
Sun SPARC Enterprise T5140 Server - Version All Versions to All Versions [Release All Releases]
Sun SPARC Enterprise T5440 Server - Version All Versions to All Versions [Release All Releases]
Information in this document applies to any platform.

Symptoms

SPD programming errors reported against DIMMs during host initialisation

Cause

Serial Presence Detect (SPD) programming data is interpreted as zeroes when read from DIMM SEEPROM during initialisation;

Apr 24 18:40:35 ERROR: MB/CMP0/BR0/CH0/D0/J0500: Unsupported SPD revision 0
Apr 24 18:40:35 ERROR: MB/CMP0/BR0/CH0/D0/J0500: Unsupported key 0, must be FBDIMM!
Apr 24 18:40:35 ERROR: MB/CMP0/BR0/CH0/D0/J0500: Unsupported burst length 0, must support 4 and 8
Apr 24 18:40:35 ERROR: MB/CMP0/BR0/CH0/D0/J0500: Unsupported number of ranks: 0, must be 1 or 2
Apr 24 18:40:35 ERROR: MB/CMP0/BR0/CH0/D0/J0500: Unsupported num of column bits: 9, must be 10 or 11
Apr 24 18:40:35 ERROR: MB/CMP0/BR0/CH0/D0/J0500: Unsupported num of row bits: 12, must be 13, 14, or 15

Although some parameter values are non-zero as shown above for column and row bits, the actual byte value in the SPD address map is indeed zero. Zero can be a valid value for some configuration parameters, so not all SPD settings will generate an error.

The DIMMs impacted by this is not consistent and may change between initialisation attempts. Certain failure modes will result in a failed configuration preventing the system from initialising;

Apr 24 18:40:42 FATAL: No useable memory branches.^M
Apr 24 18:40:42 FATAL: The HOST Processor has a configuration error, forcing a power-down^M
Apr 24 18:40:42 ERROR: Unsupported memory configuration^M
Apr 24 18:40:43 FATAL: No memory available^M
Apr 24 18:40:43 NOTICE: Sequencer Exit^M

However when the physical programming is manually checked from the fruid data (collected by fruid dataset snapshot) the programming is correct, indicating this is a read issue and not a hardware fault. This can occur following a power cycle, or after a SysFW upgrade.

Solution

There are two options for resolving this issue;

[OPTION#1]

1.- clear faults
2.- clear asrdb
3.- remove the power cables wait 2 min
4.- and then power on

[OPTION#2]

1. Reset the Service Processor

Log into the Service Processor as root
-> reset /SP

2. Reinstall the ILOM SysFW image if the issue persists after step#1

The recommendation is to follow the steps in OPTION#1 first as this includes a SP reset due to the A/C cycle, then reflash the ILOM image if the issue persists.

References

<BUG:15788606> - SUNBT7165167 SPD ERRORS AFTER 7.4.2 RESULT IN HOST INIT FAILURE

Attachments

This solution has no attachment