Monitoring Network Hosts

I l @ ve RuBoard

On a shared network medium, monitoring only the client and server systems is insufficient. Other network devices may be critical for network connectivity. Name servers were previously discussed; if the name service is unavailable, your application may be unable to connect to the server, because it can't determine the server's IP address. The router responsible for transferring your packets to the server's LAN segment may also be down.

In addition to network faults, performance problems could occur. Congestion on a shared LAN can prevent your data from being transmitted. Performance monitoring tools are discussed in the next section.

Other network devices, or hosts, may also be interesting to monitor, such as a network printer. This section discusses how you can ensure the correct operation of these additional network hosts.

Network Node Manager

As described earlier in this chapter, graphical network status monitors , such as Network Node Manager, can discover any IP-addressable device accessible from the management station, including routers, switches, hubs, and network printers. If you are trying to isolate a network problem, you may want to examine the entire path between the client and server. NNM is one tool that can help you do this. By drilling down from the global network submap, you can examine individual network segments.

Figure 6-8 shows the devices on one LAN segment. The devices include NT and HP-UX computers, as well as routers and printers. Colors indicate the status of each device, and are based on NNM's ability to poll the system and get a response.

Figure 6-8. NNM topology.

graphics/06fig08.gif

Resolving a connectivity problem may involve navigating through multiple network segments. You may have multiple routers in your environment, in which case you need to know the routes that your packets should be taking. This requires that you understand your system's routing tables.

netstat

Knowing the network topology is important for resolving communication problems. If hosts can communicate successfully on each side of a gateway or repeater, but not to systems on the other side, this most likely indicates a problem with the forwarding device.

The netstat “rn command can be used to show your current configuration of routes from the local system to remote networks, as shown in Listing 6-20. If you are experiencing long delays in packet transmission to remote networks, you may want to examine the routing algorithms being used on the network. Many networks use routing algorithms that minimize the number of hops that a packet must take, because that improves both the delay and amount of network bandwidth used, which in turn can boost network throughput.

The netstat “rs command can show statistics related to the system's use of remote routes, as shown in Listing 6-21. If a bad route is indicated, you should check whether your router is behaving properly.

If connectivity between two systems is a problem, check whether or not each system can reach other networks. The routing table may not have the right route to all the LANs. You may want to compare this configuration to the routing tables of a working system. If the routing information is wrong, you can use the route command to add and delete routing table entries.

Interconnect & Router Manager

Routers and bridges are the network devices that connect your LAN segments together. It is critical that these devices continue to work. You can use some of the same tools to monitor a router that you use to monitor your UNIX server. The router's status can be seen at a glance from tools such as NNM. Note that your UNIX server can also be configured to function as an IP forwarder.

Listing 6-20 Output from netstat showing routing information.

 cancan#netstat -rn Routing tables Destination   Gateway        Flags   Refs     Use  Interface  Pmtu 15.13.173.93  127.0.0.1      UH         1   27985  lo0        4608 127.0.0.1     127.0.0.1      UH         0  494200  lo0        4608 192.7.27.93   127.0.0.1      UH         0       0  lo0        4608 default       15.13.173.93   U          1  415198  lan2       1500 15.13.168.0   15.13.173.93   U          3   82816  lan2       1500 192.7.27      192.7.27.93    U          0       0  lan0       1500 cancan#

Listing 6-21 Output from netstat “rs showing routing statistics.

 # netstat -rs routing:         108 bad routing redirects         0 dynamically created routes         0 new gateways due to redirects         1 destination found unreachable         0 uses of a wildcard route         0 routes marked doubtful         0 routes cleared of being doubtful         0 redirects deleted #

Routers provide some of the same instrumentation as other computer systems. For example, the MIB-II is supported by all the major routers. If your router is behaving strangely and going up and down, you could check this behavior by monitoring the amount of time since the last reboot. This information is stored in the MIB sysUpTime field of the MIB-II.

Hewlett-Packard provides management software specifically for monitoring routers, bridges, and hubs. The HP Interconnect & Router Manager uses SNMP to monitor the status of AdvanceStack hubs, AdvanceStack bridges, and AdvanceStack routers. This product is integrated with NNM.

Other vendors also provide management solutions for their network devices. For example, many switches now provide Web interfaces to monitor device status and obtain statistics.

I l @ ve RuBoard