Replacing Motherboard in T5220 Plus or T5120 Plus requires workaround and/or firmware upgrade.

Asset ID:	1-73-1019881.1
Update Date:	2011-05-20
Keywords:

Solution Type FAB (standard) Sure

Solution 1019881.1 : Replacing Motherboard in T5220 Plus or T5120 Plus requires workaround and/or firmware upgrade.

Impact

- First Issue -

Incorrect Disk Slot Numbering, only applies to T5220 Plus 16 Disk Slot Backplane systems.

After replacing the Motherboard with one having On-Motherboard LSI Hardware RAID Controller Firmware 1.23.xx (or older), disk backplane numbering will be incorrect. Rather than the first disk slot being disk0 (HDD0) can be disk8 (HDD8) or higher and the second disk slot being disk9 (HDD9) and so on. This makes the system not bootable. See the Corrective Action section below for workaround.

- Second Issue -

Incorrect Operation of Disks' Ready To Remove LEDs applies to both T5220 Plus 16 Disk Slot Backplane systems and T5120 Plus 8 Disk Slot Backplane systems.

For the T5220 Plus, on HOST power-on the "Ready To Remove" LEDs on the faceplate of disks in disk slots 8 through 15 will be ON and remain ON, but should be off at that time. Solaris' prtdiag -v command and the SP's ALOM CLI showenvironment command outputs will indicate that the LEDs are OFF. A Ready To remove LED is suppose to go ON when the user uses the cfgadm -c unconfigure command, and OFF with the cfgadm -c configure command in the Disk Hot-Plug process. With this issue, they will respond in the inverse to those commands, ie; will go OFF for unconfigure and ON for configure.

For the T5120 Plus, at HOST power-on the Ready To Remove LEDs for all the installed disks will be OFF. However, if and when cfgadm -c unconfigure is run for any disk in slots 4 thru 7 (of 0 thru 7) the associated disk's "Ready To Remove" LED will NOT turn ON.

This impedes the use of the Disk Hot-Plug process. Powering off the HOST and removing a disk at that time is not a workaround. However, see the Corrective Action section below for fix.

Contributing Factors

Any Sun SPARC Enterprise T5220 Plus or T5120 Plus Server as described above are affected by these issues.

Motherboard FRUs, at the time of this FAB, all have SysFw less than 7.1.6.j as well as On-Motherboard LSI 1068 Hardware RAID Controller firmware at 1.23.xx.

- First Issue -

Incorrect disk slot numbering, occurs when a replacement Motherboard FRU with On-MotherBoard LSI Hardware RAID Controller firmware 1.23.xx or older is used to replace a T5220 Plus 16 Disk Backplane system's original Motherboard, and every time thereafter such a Replacement Motherboard is swapped in.

- Second Issue -

Incorrect Operation of Disks' Ready To Remove LEDs, occurs when a replacement Motherboard FRU, with SysFw 7.1.6.d or older, is used to replace the original or replacement Motherboards in a T5220 Plus 16 Disk Slot Backplane system or in a T5120 Plus 8 Disk Slot Backplane system.

Symptoms

See the Impact section above.

Root Cause

- First Issue -

Incorrect disk slot numbering, On-Motherboard LSI hardware RAID Controller's firmware, current replacement Motherboard 1.23.xx (and older) has the "persistence" feature ON. With persistence ON, the LSI Controller will not read the disk backplane config and will randomly assign a number to the first disk backplane slot.

The next release of the LSI controller firmware (estimated will be used as of mid-January in T5120/5220Plus Motherboards for new system builds and Motherboard replacement FRUs) will have the "persistence" feature OFF to avoid this issue. The resolution of Bug 6769499 will bring in the new On-Motherboard LSI Hardware RAID Controller firmware.

Until that new LSI Hardware RAID Controller firmware is in use, Manufacturing will be performing the Ok Prompt commands workaround, presented in the Corrective Action section below, as part of each T5220 Plus 16 Disk Slot Backplane new system build. Until T5120/5220 Plus Motherboard Replacement FRU stock is reworked to include the newer LSI Hardware RAID Controller firmware, the Ok Prompt commands workaround will need to be performed immediately after Motherboard replacement in a T5220 Plus 16 Disk Slot Backplane system.

- Second Issue -

Incorrect Operation of Disks' Ready To Remove LEDs - ILOM firware in SysFw 7.1.7.c (or older) did not fix certain Bugs which get in the way of the SysFw properly understanding the expansion sections of the T5120 Plus 8 Disk Slot Backplane (disk slots 4 thru 7) and T5220 Plus 16 Disk Slot Backplane (disk slots 8 thru 15).

All T5120 Plus 8 Disk Slot Backplane or T5220 Plus 16 Disk Backplane Systems will be shipped with SysFw 7.1.6.j installed.

Corrective Action

Although this FAB is Categorized as "Reactive", the Workaround and the Final Resolution for the first issue, as well as the Final Resolution for the second issue, are to be applied as part of the Motherboard replacement process.

Workaround:

- For First Issue - Incorrect Disk Slot Numbering:

Immediately after Motherboard replacement in an affected T5220 Plus 16 Disk Slot Backplane system, and after completing all necessary SP configurations, bring the HOST up to the Ok prompt and perform the following commands:

    setenv fcode-debug? true
    reset-all
    select <sas controller PCI path> <<< NEED THE PATH !!!!!!
    clear-persistent-all
    setenv fcode-debug? false
    reset-all << POST will run if trigger includes "user reset"
                   << Will then auto-boot if auto-boot? is true

Then continue with HOST bringup.

This workaround clears the On-Motherboard LSI Hardware RAID Controller's persist table. It does NOT turn OFF the Controller's Persistence feature.

- For Second Issue - Incorrect Operation of Disks' Ready To Remove LEDs:

See Resolution section below.

Resolution:

- For the First Issue - Incorrect Disk Slot Numbering:

The Final Resolution will be a Patch installable under Solaris CD Boot. That patch will upgrade the On-Motherboard LSI Hardware RAID Controller firmware. It is expected that Patch should be available Mid-January of 2009 or soon thereafter.

- For Second Issue - Incorrect Operation of Disks' Ready To Remove LEDs:

If the customer wishes to stay with ILOM 2.0 Based SysFw, immediately after replacing the Motherboard in a T5120 Plus 8 Disk Slot Backplane or T5220 Plus 16 Disk Slot Backplane system with a Motherboard replacement FRU having SysFw older than 7.1.7.f, upgrade the system to SysFw 7.1.7.f. Available as of February 9, 2009 via Patch 136932-08.

If the customer is ready to move to ILOM 3.0 Based SysFw, then immediately after replacing the Motherboard in a T5120 Plus 8 Disk Slot Backplane or T5220 Plus 16 Disk Slot Backplane system with a Motherboard replacement FRU, upgrade the SysFw to 7.2 minimum via Patch 139439-01, available as of December 23, 2008.

Identification of Affected Parts (how to):

- For First Issue - Incorrect Disk Slot Numbering:

It will not be until the end of February or early March 2009, that T5120/5220 Plus New System Builds and Motherboard replacement FRUs will the newer On-Motherboard LSI Hardware RAID Controller firmware.

Therefore, until this FAB is updated and republished to inform you that T5120/5220 Plus New System Builds and Motherboard replacement FRUs contain the newer On-Motherboard LSI Hardware RAID Controller firmware, be prepared to perform the "Incorrect Disk Slot Numbering" Workaround as presented in the Corrective Action / Workaround section above.

- For Second Issue - Incorrect Operation of Disks' Ready To Remove LEDs:

Before you install the replacement Motherboard, determine the SysFw version of the system by use of the SP's ALOM CLI showplatform, showhost and showsc commands.

Once you install the replacement Motherboard, use the SP's ALOM CLI showplatform, showhost and showsc commands to determine the then current SysFw version.

If the replacement Motherboard's SysFw is older than 7.1.7.f, and the customer wants to use ILOM 2.0 SysFw, upgrade the SysFw to 136932-08 7.1.7.f. If the customer wants to use ILOM 3.0 Based SysFw, upgrade the SysFw to 139439-01 7.2.

If the replacement Motherboard is at 7.2, and the customer wants the system to use ILOM 2.0 Based SysFw instead, you can back rev it to 7.1.7.f.

If you have a current explorer, the Tx000 dir of the explorer output will contain files which will tell you the SysFw version and Motherboard FRU part number / dash level. An explorer run done prior to Motherboard replacement will contain the info for the to-be-replaced Motherboard. After replacing the Motherboard and you have it at the desired SysFw level, run POST and boot the system into Solaris and then run explorer - use at least Explorer version 5.10. That explorer output will have the info for the replacement Motherboard. Inform the customer of the Solaris filesystem location of the explore output.

Comments

For more information on the Huron-Plus product reference the below URL;

http://panacea.uk.oracle.com/twiki/bin/view/Products/ProdInfoSunFireT5120

References:

BugID: 6769499, 6772715, 6769454, 6784525, 6772707, 6792056
Resolution Patches: 136932-06

For information about FAB documents, its release processes, implementation strategies and billing information, go to the following URL:

http://tns.central/fab

In addition to the above you may email:

[email protected]

Internal Contributor/submitter
[email protected], [email protected]

Internal Eng Responsible Engineer
[email protected], [email protected] Responsible Manager: [email protected], [email protected]

Internal Services Knowledge Engineer
[email protected]

Internal Eng Business Unit Group
SSG SVS (SPARC Volume Systems, Horizontal Systems)

Internal Sun Alert & FAB Admin Info
17-Dec-2008: completed draft and sent to Extended Review.
19-Dec-2008: incorporated all feedback, sending to Publish.
13-Feb-2009: Significant mods to Root Cause, Resolution and How To sections. Exact mods on file with KE.
24-Nov-2009: Corrected Product Name to swoRDFish inconsistency.

Attachments

This solution has no attachment