- Problem Statement:

After inserting a new SB to a platform, post fails with "frad
access of SEEPROM failed" and deconfigures the board in question:

FAIL Slot SB*: FRAD access of SEEPROM failed.

	* := new inserted SB, can be any slot0 [0..17]

- Symptoms:

Post fails with a failure message similar to the one below:

..
stage cpu_probe: CPU Module probe...
FAIL Slot SB1: FRAD access of SEEPROM failed.
Primary service FRU is Slot SB1.
stage cdc_probe: CDC DIMM probe...
...
                        

- Troubleshooting:

Check command line post output while running post or check postlog files 
stored under $SMSVAR/adm/[A..R]/post to verify if a platform hit this 
problem.

You can also confirm that picld is in a degraded state by checking its
status in the process table:

ps -ef | grep picld

picld -l -r            

- Resolution:

The daemon picld intermittently receives a "Segmentation Fault" and restarts
in degraded mode. "ps -ef" shows picld is running with option "picld -l -r".
picld is no longer functional. Why and where the signal is coming from is yet 
to be determined. There is an SMS plugin library in the Solaris picld. It is 
responsible for refreshing the cached information in FRAD. Because of the fact
that picld is in a degraded state (no longer functional) frad can not communicate 
to picld. So frad did NOT send the System Board REMOVAL event to picld and frad
still cached the information of the original System Board even after it has been
replaced. This is why HPOST fails.

PTS is in the process of determining why and/or where picld is receiving
the "Segmentation Fault" (BUG# 4779783). Until the root cause is identified,
we recommend implementing the current workaround.      

Problem workaround:
-------------------------

The current workaround is to restart picld and SMS

  o /etc/init.d/sms   stop
  o /etc/init.d/picld stop

  o /etc/init.d/sms   start
  o /etc/init.d/picld start


Restarting SMS will fix the HPOST failure because of the fact that by 
restarting SMS the new frad loads the cache with the data seeprom of 
the replacement System Board. Restarting picld will allow frad and picld
to communicate and function properly.

All Corrective Actions must be tested and patches must be available prior
to FIN draft submission.
N/A
        
- References and bug IDs

Esc. 541683 & Bug 4779783 have been opened to address this issue.

- Additional background information:

N/A

- Meta-Data/Problem categorization:

Product/Platform: SF12K/SF15K
Category:

- Keywords

FRAD, SEEPROM