SUMMARY: eeprom corruption

From: Nick Batchelor <Nick.Batchelor_at_unilever.com>
Date: Wed, 18 Mar 1998 12:02:59 +0000 (Z)

     Dear All
     
     Many thanks to all those who replied to my original
     question/observation which was:
     
     Having just upgraded to V5 firmware on an 8400 I started getting the
     following messages on rebooting - file open failed from eeprom0,
     eeprom0 not updated. This was fixed by running set cpu <n> ; build -e
     for each cpu in turn then putting resetting the eeprom values
     manually. Has anyone else seen the same or knows why it happens?
     
     A rather worryingly large number of people had seen it happen
     especially on 8000 series boxes. However all reports say that running
     the build -e command fixes it fine. Of course a BIG recommendation is
     to have a print out of all the eeprom settings to hand before doing a
     firmware upgrade.
     
     A number of people said that they rebuild their eeproms and nvram
     before doing an upgrade, just in case. The most detailed description
     of this came from Rodrigo Poblete and is included at the end of this
     summary.
     
     For the cause of the problem, the wisest answer came from
     paukit_at_hbsi.com who said: "As to why it happens, why does it always
     rain after you wash your car?" A more boringly technical one came
     again from Rodrigo:
     
     "I assume that the eeprom and the nvr contain data that is updated
     during the operation of the machine and that prior to a shutdown some
     status registers are changed. an interruption of the normal operation
     of the machine
     would cause the cheksum counters to not be updated, thus causing an
     inconsistency
     of the nvr or the eeprom, or both."
     
     Thanks again to Rodrigo Poblete, Marshall Spezman, Steve Thomson, Ron
     Bowman, paukit_at_hbsi.com and Karl Majer. Rodrigo's reformat everything
     solution follows:
     
     Nick
     
     ################################################################
     HI Nick
     
     I´ve experienced many of those "mysterious" corruptions of eeprom
     data.
     
     There is one recommendation:
     
     Before the upgrade procedure is executed verify the consistency of
     each of the CPU module´s NVR and EEPROM by issuing the following
     commands:
     
     first verify #s of cpus:
     P00>>sho config
     ....................... kn7cc-ab0
     ........................kn7cc-ab1
     ........................kn7cc-ab2
     2nd verify nvram
     p00>>build -n kn7cc-ab0
     3rd verify eeprom
     p00>>build -e kn7cc-ab0
     this has to be applied to each of the cpu modules.
     
     and then proceed to do the firmware upgrade.
     
     when the upgrade is finished REPEAT all three steps AGAIN.
     
     ONE WORD of CAUTION: have a copy of the
     parameters boot parameters
     (bootdef_Dev, boot_osflags, etc)
     as they are erased during the upgrade procedure)
     
     I have also experienced the corruption of the eeprom during a system
     crash or power
     interrrpution.
     this can be noticed when you initialize the machine, you get a message
     that says
     this is displayed in the last part of the initialization messages
     
     ...CPU0 has TST disabled for NVR
     
     The above procedures apply to solve this problem.
     #############################################################
Received on Wed Mar 18 1998 - 13:03:26 NZST

This archive was generated by hypermail 2.4.0 : Wed Nov 08 2023 - 11:53:37 NZDT