wiki:Apollo SM troubleshooting

Version 7 (modified by Dan Gastler, 4 years ago) (diff)

--


Eth0 not coming up:

Depending on what the Zynq RJ45 jack is connected to, the autoneg can do some strange things. Please check /fw/startup.sh to see the state of the line

root@zynq_os:~# cat /fw/startup.sh 
echo Startup script!
ethtool -s eth0 speed 100 duplex full autoneg off

Try toggling if the ethtool line is commented out.



IPMC/Zynq BOOT handshaking. Useful commands for debugging

  • SHM is the shelf manager ip
  • BLAD_IPMB_ADDR is blade IPMC address 0x9a for physical slot 1

Find blade address

ipmitool -I lanplus -H SHM -P "" fru list

Connect to Zynq via SOL (works by default, IPMC can override destination)

ipmitool -I lan -H SHM -P "" -t BLADE_IPMB_ADDR SOL activate

Force reset the blade

ipmitool -I lan -H SHM -P "" -t BLADE_IPMB_ADDR mc reset cold

The ZYNQ must write a '1' to SLAVE_I2C.S1.SM.STATUS.DONE If this doesn't happen, the IPMC will shut down the blade.


CentOS upgrades to partition 2 (root) Network:

  • Locally extract as root the tarball (tar -zxpf) and then rsync to host
  • sudo rsync -r -l -p -o -g -W --numeric-ids ./ root@host:/