Hi everyone!
Here’s a case that two many bit errors are generated at an ETH port.
Fault Description
An alarm indicating that an ETH port has too many bit errors is reported on a storage device.
Symptom
Cause
The network adapter of the storage device discards the packets whose size is larger than its MTU, resulting in frame loss.
Note: An MTU indicates the maximum length of IP packets that can be received by the network adapter.
Identification Method
1.Confirm the port ID based on the alarm information.
The internal alarm ID "CTE0.A.IOM0.P0" is used as an example.
2.Query the corresponding port ID in the configuration file in log_controller_XX_MAIN\Config.
As shown in the following figure, it is obvious that frame loss occurs, and the IP address of the port is 10.23.2.10.
Confirm that the errored frame type is frame based on the logs of the controller where the port is located.
In this example, the port is located in controller 0A.
Query the IP address 10.23.2.10 in log_controller_0A\Other\ifconfig.
The query result shows that the errored frame type of all error statistics is frame.
It can be determined frame loss occurs at the port. Besides, the MTU of the port is 1500 bytes.
If the preceding conditions are met, it can be determined that the port discards oversized frames.
Solutions
Query the port used by the host to connect to the alarm port of the storage device to check whether the MTU of the host port is larger than the MTU of the storage device port.
(1) Query method
For a Windows host, query the configuration of the network adapter to obtain or modify the value of the MTU.
For a Linux host, run the ifconfig command or view the configuration file of the network adapter to obtain or modify the value of the MTU.
(2) If the MTU of the host port is greater than the MTU of the storage device port, change them to be the same.
Note that the link between the host port and the storage device is unavailable when you modify the MTU of the host port.
Ensure that a redundant port is available for the host port when there are services on the link.
(3) If the MTU value of the host port is consistent with that of the storage device port, there is a possibility that the host sends oversized frames.
For example, the MZ512 card of an E9000 host is set to ISCSI mode,
so the E9000 host may send oversized frames periodically.
You need to set its MZ512 card to NIC mode.
Capture packets at the TX end of the host or the RX end of the storage device to check the packet size.
Check After Recovery
Run the clear port bit_error command to clear bit errors. No bit error is generated.
Note: For details about how to use restricted commands, see the Restricted Command Reference.
