Sun

Sun Enterprise 420R CPU Slave timeout

Last weekend I bought a Sun E420R that had maxed out RAM and CPU but once brought it home and powered it on everything was fine until slave CPUs 1-3 started timing out. I have already replaced the motherboard and tested each CPU in slot 0 to make sure that they all work and I still have the same problem. Could this be because the NVRAM is dead or is it another problem?

I have tried looking in the Sun manual and I couldn't find much to guide me to fixing this issue. Any help would be helpful.
So the NVRAM is dead? What do you mean with timeout?
Is it possible to get more input?
- console output and/or messages
- PROM version
- board part# and revision (type is the same used for Ultra 80 workstations)
- CPU part#'
SPARC: Sun SPARCstation 2 (1990), Sun Blade 2000 (2002)
x86: SNI PCD-3M (1989), SNI PCD-4H (1993), SNI Scenic Pro C5 (1996)
After the RAM test the 420R will display CPU Slave #3 timeout for slot 3 and the same applies for slots 1 & 2. It seems like only slot 0 is populated will the machine pass POST. I'm not at home right now but when I get back I'll post the CPU and MB part numbers.
Once before, in the first days I had also serveral problems with my Ultra 80 :?
It's because somtimes CPU modules and memory riser board are not installed properly.
A non-dead NVRAM (battery) also would be fine.

You specified cpu slot #0, but first CPU slot is #2. My misunderstanding of you text? :?
(Memory) Configuration Rules
The following memory configuration rules apply to your system:
  • At least one memory bank (all four bank slots) must be filled for the system to boot
  • Each bank used must have four identical capacity DIMMs installed (for example, four 64-Mbyte DIMMs or four 256-Mbyte DIMMs)
  • DIMMs must be added four at a time within the same bank

Note – In the Sun Enterprise 420R server, each memory bank comprises two slots on the main logic board, and two slots on the memory riser card. As you add memory, fully populate each bank by inserting DIMMs into the two slots comprising each bank on both the main logic board and on the memory riser card. For information about removing and installing DIMMs, see the Sun Enterprise 420R Server Service Manual.

DIMM capacities can differ from one bank to the next—for example, four 64-Mbyte
DIMMs installed in bank 0 and four 256-Mbyte DIMMs installed in bank 2 is permissible

(...)

(CPU) Configuration Rules
The following rules apply for your system:
  • You can install up to four UltraSPARC II CPU modules into the server.
  • The first CPU module is inserted into CPU slot 2 (J0301) ;
    two CPU modules are installed into CPU slot 2 (J0301) and into CPU slot 1 (J0201);
    three CPU modules are installed into CPU slot 2 (J0301), into CPU slot 1 (J0201), and into CPU slot 3 (J0401);
    while four CPU modules occupy all the CPU module slots.
  • If you install more than one CPU module, the modules must operate at identical clock speeds (450 MHz, for example) and they must have the same size cache memory . This generally means the CPU modules must have the same part number .
Source: Sun Enterprise 420R Server Owner's Guide, p.64 and p.66

Code: Select all

[part 1/2 of four RAM/memory banks on mainboard]
[part 2/2 of four RAM/memory banks on memory riser assembly]
[CPU Slot 3, install 3rd]
[CPU Slot 2, install 1st]       <--- insert first/single CPU in Slot 2 !
[CPU Slot 1, install 2nd]
[CPU Slot 0, install 4th]
[PCI slots]
Source: Sun Enterprise 420R Server Service Manual, p.69

Further breakdown: remove all but one CPU and all PCI cards. Using an serial cable an terminal
(...)
4. If the POST output contains an error message, then POST has failed.
The most probable cause for this type of failure is the main logic board. However, before replacing the main logic board you should:
  • a. Remove optional PCI cards.
  • b. Remove optional DIMMs. Leave only the four DIMMs in Bank A.
    (also refered as bank 0 and means: mem slots U0301+U0302 at riser board and mem slots U1301+U1302 at the mainboard)
  • c. Repeat POST to determine if any of these modules caused the failure.
  • d. If POST still fails, then replace the main logic board.
Source: Sun Enterprise 420R Server Service Manual, p. 207

Also looking at:
- Sun Enterprise 420R Server Service Manual, p. 189-191, error code definitions.
- Sun Enterprise 420R Server Owner's Guide, p. 115ff, Chapter 7 Diagnostics and Troubleshooting
- Sun Enterprise 420R Server: Field Information Notices (FIN)
- Sun Enterprise 420R Server: InfoDocs (e.g. 47161 CPU shroud or CPU card cage assembly may not seat properly in .../Sun Enterprise 420R/...)
SPARC: Sun SPARCstation 2 (1990), Sun Blade 2000 (2002)
x86: SNI PCD-3M (1989), SNI PCD-4H (1993), SNI Scenic Pro C5 (1996)