1. The customer reported that some services for users connected to an access device were interrupted. The live network topology was as shown in Figure 1. That is, Eth-trunk was deployed between R1 and R2. Because no GE optical ports were available, O/E converters were deployed between R1 and R2.
2. Some services were interrupted between two devices configured with Eth-trunk. The two devices were directly connected by optical fibers.
To address issue 2, Huawei performed the following operations and observed the following information:
1. Checked the links.
A link between R1 and an O/E converter failed, and all upstream traffic flowed to R2 over the other link. It was suspected that service interruption was caused by downlink traffic-related fault.
2. Logged in to R2 and found that R2 transmitted data to both links.
Because R1 and R2 were connected in manual mode and O/E converters were deployed between them, R2 could not detect that a port on R1 became down. In manual mode, a port is a member of the Eth-trunk as long as it is up. Therefore, R2 transmitted data to both links. By default, packets are forwarded in per destination mode, and therefore some services for some users were unavailable. (In per packet mode, severe packet loss would occur in services for all users.)
To address issue 2, Huawei performed the following operations and observed the following information:1. Checked the physical links and found that one fiber was broken.
If O/E converters are used, bind links in LACP mode.If devices are connected by fibers and do not support detection of single-point fiber faults, bind links in static LACP mode.