[ Pobierz całość w formacie PDF ]
." On router X, the state of its ethernet 0 interface and the keepalive setting, along withits IP address, subnet mask, broadcast address, IP proxy-arp setting, inbound andoutbound access lists, and IP helper-address setting can all be checked using the CiscoIOS show ip interface ethernet 0 exec command.Also, if router X has multipleconnectors on its ethernet 0 interface (in case it is a Cisco 4000 series router), thesetting of the media type can be checked using the show controllers ethernet 0command.You must get host A to be able to ping the IP address of router X s ethernet0 (10.1.1.1) successfully.Next, attempt to ping router X s serial 0 IP address (10.1.2.1).If host A can ping the10.1.1.1 address, but not the 10.1.2.1 address, the problem is probably due to its defaultgateway setting.If host A s default gateway is set to 10.1.1.2 (its own IP address), thenrouter X s ethernet 0 is probably configured with the no ip proxy-arp command.Also makesure that router X s serial 0 interface is in Up and has the correct IP configurations (addressand so on).178 Chapter 5: Diagnosing and Correcting Campus TCP/IP ProblemsNow make sure that host A can successfully ping the IP address of router Y s serial 0interface (10.1.2.2).If host A can ping router X s serial 0 address (10.1.2.1) but not the IPaddress of router Y s serial 0 interface (10.1.2.2), the reason is probably either an access liston router Y s serial interface or that router Y does not have a path back to host A s subnet (arouting issue).In this example, assume that routers X and Y can successfully ping eachother s serial 0 IP address.For router Y to be able to communicate with a device (such as host A) in subnet 10.1.1.0, itcould learn about that subnet through a dynamic protocol (such as RIP, IGRP, EIGRP,OSPF), through a static route, or by possessing a useful gateway of last resort.To see ifrouter Y has a path to host A s subnet, check router Y s IP routing table using the show iproute command.To check router Y s routing protocols and their settings, use the show ipprotocols command.Once you get host A to successfully ping router Y s serial 0 interface, pinging the IP addressof router Y s ethernet 0 (10.1.3.1) should be straightforward.If that doesn t work, checkrouter Y s ethernet 0 status and configuration using the show ip interface ethernet 0 execcommand.The next challenge is being able to ping host B.If you can ping router Y s ethernet 0(10.1.3.1) and not host B s address (10.1.3.2), you must check host B s address, subnetmask, and default gateway.Also, if host B uses its own address as the default gateway, makesure that ip proxy-arp is enabled on router Y s ethernet 0 interface.Now assume that you can ping host B by IP address, but you cannot Telnet into it.If Telnetserver is running on host B and at least one other host (such as router Y) can Telnet into it,it is likely that access lists on routers X and Y prevent the Telnet traffic between host A andhost B.Once you can Telnet into host B (from host A) by using host B s IP address, you can starttesting whether Telnet using host B s name (either its NetBIOS name or its FQDN) works.Host A can resolve host B s NetBIOS name to IP address using a local LMHOSTS file.Onsystems running Windows NT 4.0, the LMHOSTS file usually resides in the\winnt\system32\drivers\etc directory.Also, host A can find out host B s IP address bysending an inquiry to the WINS server (provided that the WINS server has suchinformation).If both of host A and host B are clients to the WINS server and cancommunicate with it successfully, NetBIOS-name-to-IP-address conversion becomes asmooth operation.In addition, host A can resolve host B s FQDN to host B s IP address byusing the HOSTS file (which resides in the same location as the LMHOSTS file but is usedfor FQDN names, which unlike NetBIOS names can be longer than 15 characters andcontain one or more embedded dots).Finally, if a DNS server is present, a FQDN can beresolved to an IP address by sending an inquiry to the DNS server as well (assuming thatthe DNS server itself has no trouble with resolving the FQDN to IP address for thisparticular host itself).The nbtstat command (entered from the CMD/Command window sprompt on Microsoft hosts) is used to display the local NetBIOS names (with the -n option),to display the content of the NetBIOS cache (with the -c option), and so on.This commandBrowser Issues in Microsoft Products 179is useful when you want to find out if a NetBIOS name is resolved to an incorrect IP address(using any of the above mentioned methods).Assuming that host A can Telnet into host B by name, most, though not all, of the problemsare now fixed.Remember that host A could not obtain an IP address through DHCP.If theDHCP server is up and running and has an available IP address in the appropriate scope forhost A, and host A can successfully ping the DHCP server, you have some router work todo.First of all, on router X, make sure that the ethernet 0 interface has the ip helper-address command pointing to the correct DHCP server (use the show ip interfaceEthernet 0 command) and that router X can itself ping the DHCP server s IP address.Also,check router X and Y to see if any access lists applied to their interfaces deny DHCP traffic(DHCP runs over UDP and makes use of ports 67 and 68).Finally, look in router X sconfiguration to see if ip forward-protocol udp is disabled using its no form (at least forUDP ports 67 and 68).If a router s interface is configured with the ip helper-addresscommand, but at the global configuration mode the router is configured with the no ipforward protocol UPD command, the ip helper-address command will be useless.TIP A trace from client to server used early in the game would reduce the number of stepsdramatically [ Pobierz całość w formacie PDF ]