On the device manager, system report the alarm as below:
The number of bit errors on ETH host port (controller enclosure CTE0, TOE interface module A.IOM0, port ID P0) is large. The system performance may be affected.
Through the collected logs or the output of the command "ifconfig", we can see the details about the NIC information as below:
According to the picture above, all of the errors are related with the frames. for the port itself, the MTU is 1500, then from the upper level(the host or switch), there will be lots of frame with MTU longer than 1500.
Then, Go to check the MTU settings on the host, it is normal
Then, Go to check the working mode of the NIC. The NIC type is MZ512 from Huawei E9000 server, after enter the NIC controller, we can see the setting for "Personality" is ISCSI, when the NIC is working with this mode, it will send the big frame periodically to the target.
The working mode for the NIC on the host is set as "ISCSI", which will send the big frames to the target periodically, when the storage received it, the alarm will appear.
Open the NIC controller, locate the "personality" setting as below and change it to the common "NIC" mode as below: