Friday, July 20, 2018

Case Study: B7079 cannot boot up after replace CPU and upgrade BIOS version

Scenario:

The customer has multiple servers of this model with the same configuration and he asked to change the CPUs for something more powerful.
After changing CPUs, all servers started OK, but one went down and would not boot at all. The hardware was not visibly damaged.
BMC was reachable, so they tried to update BIOS from the BMC, and now even BMC doesn't respond.


What steps do you suggest we do to get the server back up and running?


Hint:
1. BMC IP address mode will restore to DHCP mode after BIOS upgrade. So if BMC IP address is static IP address, the IP address will get new IP address by DHCP server.
2. B7079 cannot boot after BIOS upgrade, the BIOS firmware may damaged during BIOS upgrade process.


Solution:
There're two process can reflash BIOS version.
A. BMC is alive.
1. Connect BMC port of B7079 to DHCP network environment.
2. AC-ON B7079.
3. If BMC is alive, BMC heartbeat LED will blinking after one minute of AC-ON. 
4. use IP scanner to search BMC IP address.
5. Use Browser to remote login to BMC Web interface.
6. Upgrade BIOS via BMC Web interface.


B. BMC is dead.
1. AC-OFF B7079
2. Remove BIOS chip from motherboard.
3. Use BIOS programmer to reflash BIOS firmware.
4. put back BIOS chip.
5. AC-on and DC-ON B7079.
6. Boot B7079 to BIOS and change BIOS setting.
7. Reboot B7079 to DOS flash.
8. Type DOS command to reflash BMC firmware.
9. Reboot B7079.
10. Check BMC status.

C. If B7079 still cannot boot up. Apply RMA form to repair this motherboard.






No comments:

Post a Comment

How to fix gpu_burn compiler failure issue

System Environment: Ubuntu 22.04 LTS Server CUDA v12.0 GPU: RTX-4080 (driver 525.85.05) AP: GPU_Burn v1.1 Symptom: met error in make gpu_bur...