[Issue]
There is an alarm about bit errors on one port of storage.
2019-06-07 07:09 0xF01080015 Fault Major Unrecovered None FC host port (controller enclosure CTE0, -- controller A, port ID H2) has too many bit errors. The system performance may be affected.
[Principle]
As shown in the image above, the storage port detects the bit errors, which must be due to the poor quality of the hardware link. Its possible causes are among the SFPs of storage/switch/host, cables.
[Analysis and Suggestions]
Step1. Collect storage logs.
Step2. Open the file(\log_controller_0_MAIN\Config\config.txt). and then search the key words port ID, like ’.H2’. and ensure if you find the information of right port.
Notice: Please don’t replace the SFP directly just judging by the health status. Because it cann’t be usually solved by this way.
Step3. We need to pay more attention to the Value of RxPowerReal and TxPowerReal.
Situation1: the TxPowerReal value of fault port is lower than 300uW. Please do not hesitate to replace the SFP of storage directly.
Situation2: the RxPowerReal value of fault port is lower than 300uW. The root cause must be among the SFP of storage, the SFP of switch, the cable between storage and switch.
Solutions:
1. If the fault port is connected to switches from Huawei, please collect switch logs and contact R&D engineers.
2. If switch is not from Huawei, we need a crossing test. For example: if there are idle ports on storage or switch, we can use the original cable to connect the idle port and then observe for 5min to check if the bit error alarm is resumed. Additionally, you should get customer’s permission and ensure there are redundant links from all hosts to storage.
Tips: how to check if there are redundant links in upgrading evaluation reports.
Situation3: the RxPowerReal and TxPowerReal value are in normal range. Please collect networking topology information, storage logs, switch logs, host logs and then contact R&D engineer.
[Scope of application]
OceanStor V3/V5 series.
Dorado V3 series.
