Hello everyone,
Today, I'm going to introduce the discussion about Packet loss in the new HCIE lab.
Topic
1. Packet loss occurs on the network. How to troubleshoot the fault?
2. What are the possible causes of packet loss on the network?
Question 1
Packet loss occurs on the network. How to troubleshoot the fault?
Packet loss is a typical network fault. You can use three methods to locate the fault.
1. Top-down method.
2. Bottom-up method.
3. Dichotomy method.
Which method can be used in combination with the actual situation.
The first thing to confirm is whether the service is packet loss (impact of network quality) or the service is completely inaccessible.
1. Assuming that there is only a packet loss in the business, rather than the business being completely inaccessible, it proves that there should be no problem at the routing level. First, determine whether it is an application layer failure, access the business on the server (such as HTTP, open a web page on the server), if it is possible, it means that the business service is not faulty; if not, it means that the business service may be faulty, and the application layer needs to be performed Troubleshooting.
2. Secondly, packet loss occurs on any device during the end-to-end forwarding process, so you must first locate the device or link on which the packet loss phenomenon occurs.
3. After locating the packet loss failure point, perform troubleshooting. There may be multiple packet loss points. Tracert route locates the faulty device and links. If the tracert cannot locate the fault point due to the ups and downs of the service routing; you can use the hop-by-hop ping test.
4. If it is a device problem (no traffic filtering, no speed limit), then you can start with the physical layer to check whether it is a high load of the device, a performance problem; or a high utilization rate of the interface bandwidth, which causes congestion; a line card problem; Interface module problem; transmission cable problem; check interface traffic statistics and packet loss rate, transmission equipment, optical modems, and other equipment on the path should also be considered and checked.
Check whether the service routing table entries are normal, whether there is inappropriate QoS policy, and the rate is limited; or inappropriate queue scheduling mechanism; whether the device is under network attack; ACL and unreasonable configuration on the device; routing problem; loop Problem; link MTU problem.
Comprehensive judgments using a variety of methodologies to reduce the possibility of failure causes, such as comparison method, replacement method (Ping test by changing the host), elimination method, statistical analysis, and other methods.
If there is no problem with the network layer and the transport layer, you may consider the application layer, whether the server-side service settings are correct, whether the client is configured correctly, whether the client is VPN access, VPN access often has packet loss.
Question 2
What are the possible causes of packet loss on the network?
1. Equipment software level failure or software BUG; equipment/board/interface/module forwarding performance and software and hardware failures; transmission link quality; wrong cables; unreasonable MTU settings; and administratively closed interfaces.
2. The link layer protocol is inconsistent; the link-layer protocol negotiation fails; the second layer loop; the switch security mechanism.
3. Incorrect configuration of network layer routing protocol; network loop; unstable routing table; network attack.
4. The traffic is filtered by security mechanisms such as QoS, ACL, and other security devices.
5. The business application server itself fails, and the client fails.
That is all I want to share with you!
