Previous | Table of Contents | Next |
Step 3: Ping the IP Address of Another Workstation on the Same Segment
If you can ping the workstation, see if you can get to another workstation on the same segment. If you can, it will mean that the workstations involved are functioning correctly on the data link level; that is, theyre able to make local calls. If not, you might be dealing with a data link problem.
Dont know the address of a station on the same segment? Try pinging the broadcast address (a special address that tells all stations on the segment to respond).
The broadcast address is typically your IP network number with a 255 tacked on for the node number. Therefore, if your IP address is 192.168.10.5, with a net mask of 255.255.255.0, your node address of 5 would be replaced with 255, giving you 192.168.10.255 as a broadcast address.Some routers or UNIX workstations will allow you to ping 255.255.255.255 and will figure out the broadcast address for you, but Windows will not.
Once you ping the broadcast address, open the Address Resolution Protocol (ARP) table (this gives you information about IP-to-MAC address resolution) by typing this:
arp -a
If you see the following message, youve probably got a problem:
No ARP Entries Found
If other folks on this segment are okay, check the workstation cable and the card. Otherwise, the whole shebang will look something like this:
C:\>ping 167.195.163.255
Pinging 167.195.163.255 with 32 bytes of dat A Reply from 167.195.163.255: bytes=32 time<10ms TTL=128 Reply from 167.195.163.255: bytes=32 time<10ms TTL=128 Reply from 167.195.163.255: bytes=32 time<10ms TTL=128 Reply from 167.195.163.255: bytes=32 time<10ms TTL=128
C:\> arp -a
Interface: 167.195.163.7 on Interface 1 Internet Address Physical Address Type 167.195.163.3 00-00-c9-0b-ec-7f dynamic 167.195.163.9 00-00-c9-14-93-17 dynamic 167.195.163.15 00-00-c9-1e-30-97 dynamic 167.195.163.17 00-05-24-dd-79-ea dynamic
This technique does two things for you: It gives you a handy list of MAC-to-IP addresses, and it tells you that this network card, in conjunction with the TCP/IP stack, is working fine.
Take notes about what works and what doesnt. Youll begin to see how the picture starts to come together.
Step 4: Ping the Segments Router
Try pinging the segments router. Chances are, if you were able to ping other stations on the segment, youll be able to ping this, too.
Step 5: Ping the Server by Name and IP Address
Ping the server by IP address and then try to ping it by name. The important thing to remember is that you want to troubleshoot by IP number before bringing name services into the picture; otherwise, you might confuse an already complex issue.
If pinging the IP address works but pinging the name doesnt, you should investigate the DNS configuration of the workstation or check the DNS server itself.
You can check the DNS configuration with the nslookup tool in NT or UNIX. (More on DNS in Hour 19, Internet/Intranet Troubleshooting.)
If pinging the IP address doesnt work, youll definitely want to traceroute the address. Listing 17.1 shows a healthy traceroute from a good workstation to a server (see Figure 17.2).
Figure 17.2 Using the traceroute command is helpful when youre troubleshooting routing problems.
The traceroute command on most routers and UNIX systems is actually spelled out (traceroute), whereas Windows NT/95s traceroute command is typed as tracert.
Listing 17.1 A Healthy Traceroute
C:\>tracert 167.195.165.15
Tracing route to mail2.blibdoolpoolp.com [167.195.165.15] over a maximum of 30 hops: 1 10 ms 10 ms 10 ms 167.195.163.1 2 60 ms 70 ms 61 ms 167.195.174.2 3 60 ms 71 ms 60 ms mail2.blibdoolpoolp.com [167.195.165.15] Trace complete.
If the only router serving a subnet is down, you might get a destination unreachable message, depending on your routing setup. Some implementations of traceroute will show this as !N.This happens because the router isnt alive; therefore, it cant generate the routing protocols that advertise the network to other routers. If no router knows about it and no default route passes it off to another better informed router, then the destination probably doesnt exist.
Each and every router involved is shown in the traceroute output. It goes without saying that if you see the traceroute just stop, then youve discovered the most likely point of failure. For example, if the traceroute in Listing 17.1 had stopped at 167.195.163.1 and never made it to the next hop (167.195.174.2), its likely that theres a problem with the wide-area link (maybe the telephone company) or router 1 is down. A successful ping of 167.195.174.1 and an unsuccessful ping of 167.195.174.2 would verify this. (It would also show that router 2 was doing its job but was unable to contact the other side of the wide-area link.)
Bear in mind that only one interface from each router will be shown on a traceroute. Notice that theres no record of the packet passing through 167.195.165.1 or 167.195.174.1.
You can also spot misconfigurations with the traceroute command. For example, if you try to traceroute the mail server and get the following output, then router 3 is seriously confused, either by bad configuration information or incorrect information from a routing protocol:
C:\>tracert 167.195.165.15 Tracing route to mail2.traceroute.com [167.195.165.15] over a maximum of 30 hops 1 <10 ms <10 ms <10 ms 167.195.163.2 2 <10 ms 10 ms 10 ms 167.195.194.1 3 <10 ms 10 ms 10 ms 167.195.163.2 4 10 ms 10 ms 10 ms 167.195.194.1 5 10 ms 10 ms 20 ms 167.195.163.2 6 10 ms 20 ms 20 ms 167.195.194.1 7 10 ms 20 ms 20 ms 167.195.163.2 ^C
Here, router 4 says, Hey, this isnt a packet for me, take it back! Then router 3 says, Duh, no, George, this is a packet for you! In this case, youd have to look at several routers routing tables to figure out what the deal was. (This is called a routing loop, which typically comes about when a router is configured with an incorrect static route.)
Previous | Table of Contents | Next |