Huawei DHCP Dial Up Failed After MA5800 OLT Reboot

DHCP dial up failed after MA5800 OLT reboot

Issue Description

many users of the MA5800 site in the live network reported DHCP dial-up failure and cannot obtain IP address.

Handling Process

1. When the issue occurred, performed remote capture for the faulty user 0/1/8 ONT2 and analyzed the packets. found that sometimes the ONT sends a DHCP Discover but does not receive an Offer. Sometimes, after the ONT sent the DHCP Discover packet, it needed to wait 17 seconds to receive the DHCP Offer packet, causing the ONT waiting timeout.

2. Use the command "(diagnose)%%display cpu-overload-control queue-config" to query the OLC(Overload Control) message queue statistics. found that many DHCP Discover are congested in the OLC queue 8 and packet loss occurs.

3. Use the command "display udm user number" to query the number of users in the connecting state, which is 3027. This means that 3027 DHCP users are initiating DHCP dialup at the same time. As a result, the OLT receives a large number of DHCP Discover/Request packets and DHCP Offer/ACK packets from the DHCP server.

4. Finally, power off all the service boards and then power on them one by one. The services are restored.

5. Analyzing the log and debug log, in the current R20C10SPH302 version, when dealing with a scenario where a large number of DHCP users dial up at the same time, may experience OLC queue congestion and be unable to process DHCP messages in a timely manner, resulting in Discover packet loss or slow forwarding of DHCP Offer/ACK to end user.

Root Cause

During DHCP dialup, message exchange between processes is introduced. As a result, the DHCP dialup processing time increases. When a large number of users perform DHCP dialup concurrently, a large number of dialup packets may time out or the packets were discarded, causing a dialup failure.

Solution

Temporary solution：

Power off all the service boards and then power on them one by one. The services are restored.

Services can be restored by powering off and then powering on the boards one by one. This reduces the number of concurrent online users and the number of DHCP packets sent to the CPU, preventing DHCP queue congestion of the OLC.

Final Solution：

The DHCP dialup process is optimized in R20C10SPH310, so that dialup packets are processed as soon as possible and the user online rate is increased.

Suggestions

It is recommended to install the latest patch version which could resolve issues.