SGI: Hardware

SGI 0300 NVRAM issue

Hi,

My o300 seems to be playing up. Upon boot I get the error

NVRAM chacksum is incorrect

Then sometimes the walking of the SCSI chain fails but sometimes it succeeds.

I have tried a reset form L1 but I still get the NVRAM error. Is the NVRAM something that can be replaced?

If so what is the chipset and what is the procedure for doing so?

Any help would be very much appreciated.

Cheers
apologies

obviously chacksum should read checksum - my typo!! ;)
Could you post the output of the L1 commands "serial all" and "log"?

No need to boot the system, as long as power is connected the L1 should be running.

_________________
***********************************************************************
Welcome to ARMLand - 0/0x0d00
running...(sherwood-root 0607201829)
* InfiniteReality/Reality Software, IRIX 6.5 Release *
***********************************************************************
hi,

This is the output that you rrequested

01c01-L1>serial all

Data Location Value
------------------------------ ------------ --------
Local System Serial Number NVRAM M2004389
Reference System Serial Number NVRAM M2004389
Local Brick Serial Number EEPROM MPD712
Reference Brick Serial Number NVRAM MPD712


EEPROM Product Name Serial Part Number Rev T/W
---------- -------------- ------------- -------------------- --- ------
NODE IP45_4CPU MPD712 030_1797_001 B 00
IO8 IO8 MPA366 030_1673_004 C 00

EEPROM JEDEC-SPD Info Part Number Rev Speed SGI
---------- ------------------------ ------------------ ---- ------ --------
DIMM 0 CE000000000000000C0C4700 M3 46L2820BT2-CA0 2B 10.0 N/A
DIMM 2 7F94FFFFFFFFFFFF9374D80D SM57264DSGI100C3 00FF 8.0 N/A
DIMM 1 CE000000000000000C0A4700 M3 46L2820BT2-CA0 2B 10.0 N/A
DIMM 3 7F94FFFFFFFFFFFFB374D80D SM57264DSGI100C3 00FF 8.0 N/A


001c01-L1>log
02/08/13 03:37:23 power up (PANEL)
02/08/13 03:37:30 reset again MIPS
02/08/13 03:54:11 power down (COMMAND)
02/08/13 03:55:25 SMP unregistering events
02/08/13 03:55:25 UNREG: 30005140 0 4
02/08/13 03:55:26 SMP-R: UART:UART_NO_CONNECTION
02/08/13 04:05:05 power up (PANEL)
02/08/13 04:05:13 reset again MIPS
02/08/13 04:07:12 power down (PANEL)
02/08/13 04:08:10 power up (PANEL)
02/08/13 04:08:18 reset again MIPS
02/08/13 04:08:26 SCC WR 0 (len=6) - UART:UART_TIMEOUT
02/08/13 06:19:43 power down (COMMAND)
02/08/13 06:20:11 SMP unregistering events
02/08/13 06:20:11 UNREG: 30005140 0 4
02/08/13 06:20:12 SMP-R: UART:UART_NO_CONNECTION
02/09/13 05:34:07 power up (PANEL)
02/09/13 05:34:15 reset again MIPS
02/09/13 05:34:50 power down (PANEL)
02/09/13 05:35:24 power up (PANEL)
02/09/13 05:35:32 reset again MIPS
02/09/13 05:35:39 SCC WR 0 (len=6) - UART:UART_TIMEOUT
02/09/13 05:37:31 power down (PANEL)
02/09/13 05:37:32 power up (PANEL)
02/09/13 05:37:40 reset again MIPS
02/09/13 05:41:11 power down (COMMAND)
02/09/13 05:42:04 power up (PANEL)
02/09/13 05:42:12 reset again MIPS
02/09/13 05:42:25 SCC WR 0 (len=6) - UART:UART_TIMEOUT
02/09/13 05:49:34 power down (COMMAND)
02/09/13 05:49:50 SMP unregistering events
02/09/13 05:49:50 UNREG: 30005140 0 4
02/09/13 05:49:51 SMP-R: UART:UART_NO_CONNECTION
02/13/13 03:13:18 power up (PANEL)
02/13/13 03:13:26 reset again MIPS
02/13/13 03:13:37 SCC WR 0 (len=6) - UART:UART_TIMEOUT
02/13/13 03:15:56 reset (COMMAND)
02/13/13 03:18:59 power down (COMMAND)
02/13/13 03:19:14 power up (PANEL)
02/13/13 03:19:22 reset again MIPS
02/13/13 03:22:54 power down (PANEL)
02/13/13 03:22:55 power up (PANEL)
02/13/13 03:23:03 reset again MIPS
02/13/13 03:24:54 power down (PANEL)
02/13/13 03:25:36 SMP unregistering events
02/13/13 03:25:36 UNREG: 30005140 0 4
02/13/13 03:25:37 SMP-R: UART:UART_NO_CONNECTION
02/17/13 04:02:11 power up (PANEL)
02/17/13 04:02:19 reset again MIPS
02/17/13 04:03:52 power down (PANEL)
02/17/13 04:03:54 power up (PANEL)
02/17/13 04:04:02 reset again MIPS


Hope that this gives you some insight.

Can't say that at this point in time it means much to me. Any insight you can provide would be useful.

Cheers
I looked through the forum archives and found a couple of mentions of "NVRAM checksum is incorrect" issues.

The first most closely matches your problem, but unfortunately the OP (tstover) never followed up to mention if he'd found a solution. viewtopic.php?f=3&t=16726790

The second is a similar issue but not your exact issue (or hardware), the OP does follow up (thanks pierocks!) that replacing the PROM battery resolved his NVRAM checksum errors: viewtopic.php?f=3&t=16721935t&start=15#p7309808

The third involves an O350 rather than an O300, and the OP there specifically mentions ordering a replacement battery for the I/O controller (an IO9 in the O350, while your O300 has an IO8). No specific mention that the battery resolved the NVRAM checksum error, but perhaps the OP (mia) will see this thread and offer comment. viewtopic.php?f=3&t=16727051

The ST M4T28-BR12SH1 mentioned by mia is located on the IO9 controller used in the O350/Tezro/Altix. I don't have an O300 so I can't offer any advice as to where the equivalent O300 battery might be located.

_________________
***********************************************************************
Welcome to ARMLand - 0/0x0d00
running...(sherwood-root 0607201829)
* InfiniteReality/Reality Software, IRIX 6.5 Release *
***********************************************************************
recondas wrote:
The third involves an O350 rather than an O300, and the OP there specifically mentions ordering a replacement battery for the I/O controller (an IO9 in the O350, while your O300 has an IO8). No specific mention that the battery resolved the NVRAM checksum error, but perhaps the OP (mia) will see this thread and offer comment. viewtopic.php?f=3&t=16727051

The ST M4T28-BR12SH1 mentioned by mia is located on the IO9 controller used in the O350/Tezro/Altix. I don't have an O300 so I can't offer any advice as to where the equivalent O300 battery might be located.


I managed to dig up a photo that shows the component side of an IO8 (it mounts in the O300 component side down). It appears the IO8 (in your O300) uses a battery that's at least visually similar to the ST M4T28-BR12SH1 mia mentioned replacing on the IO9 used in his O350:
mia wrote:
Today's error message are:
Code:
NVRAM checksum is incorrect: reinitializing.
I've ordered a spare battery, a ST M4T28-BR12SH1 ($4 @ mouser) ; hopefully that will work.

Replacing that four-dollar battery ended mia's NVRAM checksum errors.

Here's a photo of an IO9. The ST M4T28-BR12SH1 battery mia mentions is indicated by the arrow (the battery is yellow):
Attachment:
IO9_battery.jpg
IO9_battery.jpg [ 177.73 KiB | Viewed 289 times ]
...and here's the photo I found of an IO8, what I suspect is the battery is also yellow (and marked by an arrow):
Attachment:
IO8 Assembly.JPG
IO8 Assembly.JPG [ 77.7 KiB | Viewed 289 times ]
There's not enough resolution in the googled IO8 photo to read any part numbers, so you'll probably want to pull your IO8 and confirm the part number before you make a decision to order. For reference, here's the link to the ST M4T28-BR12SH1 offered by Mouser: http://www.mouser.com/ProductDetail/STM ... bHObpv1g==

If you pull your IO8 and find a battery with a different part number please let us know what you've found.... ditto if you replace the battery and find it fixes the NVRAM checksum is incorrect error message.

Just in case the mouser link goes missing in action over time, I've attached the ST M4T28-BR12SH1 data sheet for posterity's sake.

_________________
***********************************************************************
Welcome to ARMLand - 0/0x0d00
running...(sherwood-root 0607201829)
* InfiniteReality/Reality Software, IRIX 6.5 Release *
***********************************************************************
As recondas pointed out, I bought a M4T28-BR12SH1 to fix a similar issue I was getting on a Origin 350, and it worked. Odds are, it will work for you too.

_________________
:Onyx2:
Hi,

thanks to all who have relied. I pulled my IO8 and found the same yellow battery with the same part number on my o300.

I am going to go with that first as it seems to have fixed most people's issues.

There is another NVRAM chip also - a DALLAS DS1742W-120

Any idea why there are two and when they are used. I am guessing that they are used by the L1/L2 and/or the system at different points.

I'll post back when I have my new yellow battery.

thanks again.

Cheers