SunFire v245 Environmental PIC micro controller fix

Wez · February 24, 2010, 10:52am

Hello All,

We have a v245 that powered itself off for no reason. Searching around we discovered there is a bug were a temp error is being reported but doesnt actually exist, SUN released an EPIC update, the README is shown below but even after upgrading the ALOM to the latest v1.6.10 we do not have the commands required.

Any ideas?

Patch-ID# 139735-01

NOTE:
***********************************************************************
READ THE TERMS OF THE AGREEMENT ("AGREEMENT") IN THE LEGAL_LICENSE.TXT
FILE CAREFULLY BEFORE USING THIS SOFTWARE. BY USING THE SOFTWARE, YOU
AGREE TO THE TERMS OF THIS AGREEMENT. IF YOU DO NOT AGREE TO ALL OF THE
TERMS, PROMPTLY DESTROY THE UNUSED SOFTWARE.
***********************************************************************

Keywords: sunfire v215 v245 server random power off erroneous overtemp
Synopsis: random power off with MB.OVERTEMP LED ON but NO ALOM/Solaris event log
Date: Nov/17/2008

Install Requirements: Additional instructions may be listed below

Solaris Release: 9 10

SunOS Release: 5.9 5.10

Unbundled Product: Environmental PIC micro controller - PIC18F67J10

Unbundled Release: EPIC Version 0x5b

Xref:

Topic:

Relevant Architectures: sparc
NOTE:
Fixed in EPIC Version 0x5b

BugId's fixed with this patch: 6662423

Changes incorporated in this version: 6662423

Patches accumulated and obsoleted by this patch:

Patches which conflict with this patch:

Patches required with this patch:

Obsoleted by:

Files included with this patch:

copyright
main_R20081009_f8a1_v5b.HEX
README.139735-01

Problem Description:

6662423 random power off with MB.OVERTEMP LED ON but NO ALOM/solaris event log

Patch Installation Instructions:

ABOUT THIS PATCH
----------------

IMPORTANT: PLEASE NOTE THAT THIS PATCH IS FOR SUNFIRE V215 V245 SERVERS ONLY.

Instructions to update EPIC

A) Connect and configure the service processor network management port.

Connect a network cable to the port labeled "NET MGT<-->" on the rear of the
server.

Connect the other end of cable to a switch in your network.

Assign an IP address to the network management port via DHCP or static address
allocation.

B) Set the ALOM ssqamode true

Use showsc to view netsc_enetaddr and sys_hostname:

sc> showsc
...
netsc_enetaddr 00:14:4f:e9:06:94
sys_hostname wgs99-143

Combine the last three characters on the sys_hostname with
the second digits of the last bytes of netsc_enetaddr:

wgs99-143
^^^

00:14:4f:e9:06:94
^ ^ ^

The password in this case would be "143964"

Set sc_ssqamode true:

sc> setsc sc_ssqamode true <password>

C) Set the ALOM servicemode true

sc> setsc sc_servicemode true

D) Download the new EPIC binary

Check version before download by reading the appropriate EPIC register

sc> peek 0xb4 1 5
0x5a (this is the shipping version of EPIC - version 5a)

Place the .HEX file in an accessible directory. The epic update command uses FTP
to download the file.

Put system into standby mode

Run the epicupdate command

sc> epicupdate -s ab.cd.ef.gh -f /home/path/to/somewhere

where ab.cd.ef.gh is the IP address of a server containing the image file

e.g.

sc> epicupdate -s 10.6.XX.XX -f /home/V2x5/EPIC_FW/main_test_version_5b.HEX
Username: <user>
Password: ********

Image downloading...
Image flashing...
................................................................
Image verification...
EPIC update completed sucessfully.
Reset ALOM to use new EPIC image.
sc>
sc> resetsc -y

Check version after upgrading

sc> peek 0xb4 1 5
0x5b (this is the modified version of EPIC - version 5b)

E) Reset the SC parameters to default values.

Reset ssqamode to default (false)

sc> setsc sc_ssqamode false

Reset service mode to default (false)

sc> setsc sc_servicemode false

It is possible you may see an error during the Image verification after flashing the
device: e.g.

Image verification...
EPIC image verification failed. Bad record 0x4
EPIC update failed.
EPIC update failed. Please retry

This is normally a false error. You should perform resetsc -y and check the EPIC version to prove it has been updated properly.

*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*

Special Install Instructions:
-----------------------------

**************************************************************************
**WARNING:****WARNING:****WARNING:****WARNING:****WARNING:****WARNING:**

PLEASE READ THE INSTRUCTIONS CAREFULLY AND ENTIRELY. PLEASE FOLLOW THE
INSTRUCTIONS EXACTLY AS THEY ARE STATED.

FAILURE TO FOLLOW INSTRUCTIONS COULD EITHER RESULT IN A DEFECTIVE BOARD
OR UNSUCCESSFUL EPIC FIRMWARE UPDATE.

THIS IS NOT A STANDARD PATCH IN THAT IT DOES NOT USE THE INSTALLPATCH
AND BACKOUTPATCH SCRIPTS.

PLEASE REVIEW NOT ONLY THIS README BUT ALSO ADDITIONAL DOCUMENTATION
GIVEN.

**WARNING:****WARNING:****WARNING:****WARNING:****WARNING:****WARNING:**
**************************************************************************
**************************************************************************

README -- Last modified date: Friday, Nov 14, 2008

README -- Last modified date: Thursday, December 18, 2008

incredible · February 24, 2010, 11:31am

we do not have the commands required

I don't really understand what you want
From the sc> can you provide us the "showhost" output ?

Wez · February 24, 2010, 12:20pm

The sc> doesnt have the commands available needed to execute the upgrade, ie peek, epicupdate etc. It doesnt even have the showhost command you asked for

Copyright 2009 Sun Microsystems, Inc. All rights reserved.
Use is subject to license terms.

Sun(tm) Advanced Lights Out Manager 1.6.10 (**********)

Please login: admin
Please Enter password: ********

sc> showhost
Invalid command. Type 'help' for list of commands.
sc> help
Available commands
------------------
poweron [-c] {FRU}
poweroff [-y] [-f]
removefru [-y] {FRU}
reset [-y] [-x] [-c]
break [-y] [-c]
bootmode [normal|reset_nvram|diag|skip_diag|bootscript="string"]
console [-f]
consolehistory [-b lines|-e lines|-v] [-g lines] [-p|boot|run]
dumpconfig [-t] -s <IPaddr> -f <pathname>
showlogs [-b lines|-e lines] [-g lines] [-v] [-p logtype[r|p]]
setlocator [on|off]
showlocator
showenvironment
setfru -c <Customer data>
showfru [-g lines] [-s|-d] [FRU]
showplatform [-v]
setkeyswitch [-y] [normal|stby|diag|locked]
showkeyswitch
showsc [-v] [param]
shownetwork [-v]
setsc [-r [y]] [param] [value]
ssh-keygen {-l|-r} -t {rsa|dsa}
restartssh [-y]
setupsc
showdate
setdate [[mmdd]HHMM | mmddHHMM[cc]yy][.SS]
resetsc [-y]
restoreconfig [-t] [-x] [-y] -s <IPaddr> -f <pathname>
flashupdate -s <IPaddr> -f <pathname> [-v]
setdefaults [-y] [-a]
useradd <username>
userdel [-y] <username>
usershow [username]
userpassword <username>
userperm <username> [c][u][a][r]
password
showusers [-g lines]
logout
help [command]
sc>

---------- Post updated at 05:20 PM ---------- Previous update was at 04:39 PM ----------

Here is the link to the patch on the sunsolve site :-

#139735-01: random power off with MB.OVERTEMP LED ON but NO ALOM/Solaris event log

incredible · February 24, 2010, 12:44pm

resetsc, worst case, do a COLD boot of the system

Wez · February 25, 2010, 3:55am

The SC has been reset a couple of times and the server has been powered on and off, is this what you mean?

incredible · February 25, 2010, 5:01am

poweroff the system, Pull out the power cords and leave it untill a couple of minutes. Plug in power cords again and power on.

Was this issue raised to Sun Support ? The best alternative is to get the system board replaced if all types of isolation fails.

Wez · March 1, 2010, 5:18am

The server is actually working fine apart from not being apply this fix for the known problem.

Whats the best way to open a ticket with SUN now that they have been taken over by Oracle, I looked for an online ticket/service area but didnt find anything.

incredible · March 1, 2010, 9:41am

call up SUN hotline. check up your system contract by providing the system serial number

bono71 · March 31, 2010, 4:20am

Hello Wez!

I have the same problem as you have. Did you find a solution?
I did an OBP and an ALOM update on a V210 and have no command like "epicupate". At the other side i did not found any EPIC-Update-Patch for V210 on SUN (oracle) Supportsite. Maybe this exists only for V215 and V245.
Problem is I have no V245 to test it.

would be great if somebody has an idea (no I can�t contact my sunsupport - because my local sunsupport is totaly incompetent :()

regards
Mike