Question: How to Check Fibre Channel Switch Link Faults?
Answer:
1. Check whether the port link is normal.
Open the collected switch logs. If the switch logs are collected by running the supportshow command, query fabriclog –show in the command output. If the supportsave is used to collect information, find the xxxxxx-SUPPORTSHOW_ALL.txt file in the log file after the Quickcheck tool processes the information, open the file, and query the printed information of the fabriclog –show command. The fabriclog –show command output contains a large amount of information. You need to narrow down the query scope based on the time when the problem occurs. For details about Fabriclog –show, see the Fabric OS Command Reference.
Run the fabriclog –show command to view all fabric event logs on the switch. Pay attention to the online and offline events of switch ports. As shown in Figure 2-4, SCN Port Offline indicates that the physical link of the port is interrupted (with the "*Removing all nodes from port" event, all logical devices of the port are removed). According to the corresponding port item, the physical link is offline on port 5. The time is 09:59:08.822029.
Figure 2-4 Port device offline event
Note: To save the storage space, each event log printed by the switch contains only the specific time point and does not contain the date information. (If the interval between the fabric log printed by the switch and the previous fabric log is longer than 8 hours, the switch prints the current date before printing the log. To view the date information, view the latest date print record and calculate the correct date based on the corresponding time.
2. Analyze the port bit error statistics.
In the xxxxxx-SUPPORTSHOW_ALL.txt file, you can query porterrshow or portstatsshow PortNumber to view the bit error information of the corresponding port, as shown in Figure 2-5 and Figure 2-6.
The bit error statistics of a switch port are the historical bit error statistics collected after the port goes online. The existing port bit error statistics cannot determine whether the current port is faulty. The best way to locate the fault is to clear the port bit error statistics on the switch. After one to two hours, collect switch logs again to check the port bit error statistics. After bit errors were cleared, a large number of bit errors were still collected on the port. Therefore, the physical link was abnormal. Run the portstatsclear PortNumber command to clear the statistics (including the bit error statistics) of the corresponding port.
Figure 2-5 Port bit error information displayed by porterrshow
Figure 2-6 Checking the port bit error information
3. Check the transmit and receive power of the optical module.
In the xxxxxx-SUPPORTSHOW_ALL.txt file, run the sfpshow –all command to query the information about optical modules (including the transmit and receive power) on all ports. The command output contains the current transmit and receive power of the optical module and the corresponding alarm threshold, check whether the optical module is normal. as shown in Figure 2-7.
If the transmit optical power of an optical module on a port is too low or too high, the optical module is abnormal. Replace the optical module on the port.
If the receive optical power of the optical module is too low, the possible causes are as follows:
1. No optical fiber is connected to the port.
2. The peer device does not transmit optical signals or the transmit power of the peer optical module is too low.
3. The optical fiber is aged and the light sensitivity of the link is poor.
Check the links based on the three possible causes listed in, replace the faulty component, and observe the fault.
Figure 2-7 Optical module information
