1. This network used USG2160 before, operated normally. After replacing USG5120, client always fail to access the internet. LAN is constructed by protocol convert-electronic to light to access the USG5120 G0/0/0 port, bandwidth is 100M, duplex, when fail to access internet, G0/0/0 port’s physical protocol are all UP, and learnt the opposite port ARP, but can’t ping through the opposite port IP, can only recover the network by using shutdown undoshutdown interface.
2. Confirm the problem with client, make sure the G0/0/0 port use self-negotiate protocol at the beginning, after the problem emerge, force duplex, problem still unsolved.
3. Protocol converter brand: Silver camel ring.
1. Recover client's operation quickly, specific operation:
Delete the GE0/0/0 Port's configuration, set the speed and duplex to self-negotiation mode as follows:
2. Checke client's equipment configuration, no problem.
3. Version is too old, update to newest version（v100r005c00spc500）,but the problem still exists.
4. Catch package to analyze. By using long distance catching package, analyze the equipment.
Analyze the package information and log information, during the catching, only get the package that sending off from G0/0/0, didn't get any receiving package. But because long distance catching package belongs to software catching package, it can only prove the package was sent on the software level, but can't confirm whether understructure driver and physical port sent the package.
5. Analyze the understructure driver and physical port by catching package as follow steps:
(1) When problem emerge, please clear interface package count.
reset counters interface
(2) Execute ping gateway 184.108.40.206.
(3) Check the interface count content again.
dis interface GigabitEthernet
(4) After execute these operations, enter the hidden view to check data stat information.
(5) Clear the data stat information
[USG2200-hidecmd]display dataplane statistic module
(6) Execute ping gateway 220.127.116.11
Check the data stat information
USG2200-hidecmd]reset dataplane statistic module
Finally collect interface package stat information again
6. Through analysis: The problem possibility of our equipment is very little, the problem should be on the switch that directly connected to our equipment. The cause should be the ARP list didn't correctly refresh or there was other same ip address(18.104.22.168)
7. By contacting with the carrier’s engineer, found that the cause of the problem is the public ip address clash. After change the public IP address, problem didn't exist.
1. Client's equipment configuration problem.
2. Client's equipment version is too old.
3. Other company equipment's problem.