Symptom:
The server fails to start up and does not automatically enter the OS. The server is suspended in the BIOS phase. The system is recovering data, please wait... is displayed on the BMC KVM screen, as shown in the following figure.
Diagnosis:
There are dust, oil stains, or abnormal small particles around the memory module, which affect the contact factor of the memory module.
The BIOS version is 1.25 or later.
A server is suspended during startup and cannot be recovered by powering it off.
Cuase:
For BIOS versions later than 1.25, the Taishan 2280 V2 and 5280 V2 servers use the Cold Boot Fast process by default for cold startup. If a physical fault occurs on the memory module when a server in powered off, the Cold Boot Fast process is triggered after the server is powered on. The faulty memory module cannot be identified and is not isolated. As a result, the server is suspended during startup.
Solution:
Disconnect the AC power supply and open the chassis of the faulty node.
Exchange the positions of the two memory modules.
After the chassis is installed, restore the power supply and power on the server.
After the server is started, log in to the BMC. View the current alarms and events and locate the faulty memory module.
Power off the server and replace the faulty memory module.
