Many people are not aware of how to make the best use of the features the the Virtual PCs (
VPCs) program, so I thought I’d run a small tutorial.
Technical note: This tutorial was written using VPCs version 0.21 –
and some of the functions described here will not work exactly the same
on 0.20a, and for version 0.16 – you will have to use echo instead of
TCP and
UDP ping options.
To get started, I’ll assume you are working with the
GNS3 Workbench,
and have access to some working configurations. To keep things simple,
I’m going to start with the nice triangular network that is found in
Wendell Odom’s
CCNA books. (If you are not working with the
GNS3 Workbench, then download the topology, config and startup.vpc files
here). Or of course, you could just build your own network based on the diagram below.
To get to the working example, start by double-clicking the
GNS3 Files-More exercises here icon. This will open up a folder of exercises, and from there open the exercise
3Router-ICND_Book_Example. In this folder you will see several files whose name begin with
startup… Find the file called
startup.2 (eigrp configured) and double click on it. You will be prompted with:
Do you want to run “startup.2 (eigrp configured)”, or display its contents?”
Click on
Run and the lab will load.
In the GNS3 application, from the menu select
Control->Start/Resume all devices to start your routers, and once the routers have started, from the menu select
Control->Console to all devices icon. After all the routers have started, and the routing protocol converged, you are ready to start learning about VPCs.
Start by activating the VPCs Konsole window. You should see:
Executing the startup file
PC1 : 10.1.1.1 255.255.255.0 gateway 10.1.1.251
PC2 : 10.1.2.1 255.255.255.0 gateway 10.1.2.252
PC3 : 10.1.3.1 255.255.255.0 gateway 10.1.3.253
NAME IP/CIDR GATEWAY MAC LPORT RPORT
VPCS1 10.1.1.1/24 10.1.1.251 00:50:79:66:68:00 20000 30000
fe80::2050:79ff:fe66:6800/64
VPCS2 10.1.2.1/24 10.1.2.252 00:50:79:66:68:01 20001 30001
fe80::2050:79ff:fe66:6801/64
VPCS3 10.1.3.1/24 10.1.3.253 00:50:79:66:68:02 20002 30002
fe80::2050:79ff:fe66:6802/64
VPCS4 0.0.0.0/0 0.0.0.0 00:50:79:66:68:03 20003 30003
fe80::2050:79ff:fe66:6803/64
VPCS5 0.0.0.0/0 0.0.0.0 00:50:79:66:68:04 20004 30004
fe80::2050:79ff:fe66:6804/64
VPCS6 0.0.0.0/0 0.0.0.0 00:50:79:66:68:05 20005 30005
fe80::2050:79ff:fe66:6805/64
VPCS7 0.0.0.0/0 0.0.0.0 00:50:79:66:68:06 20006 30006
fe80::2050:79ff:fe66:6806/64
VPCS8 0.0.0.0/0 0.0.0.0 00:50:79:66:68:07 20007 30007
fe80::2050:79ff:fe66:6807/64
VPCS9 0.0.0.0/0 0.0.0.0 00:50:79:66:68:08 20008 30008
fe80::2050:79ff:fe66:6808/64
VPCS[1]>
The
[1] shows that your focus is Virtual PC #1. From the output above, you can see that
VPCS1‘s IP address is
10.1.1.1 and its
default gateway is
10.1.1.251. In the GNS3 topology, it is marked as PC1Bugs.
Lesson #1 – Successful pings
In this lesson, you will see:
- the purpose of the startup.vpc file
- how to check your configuration
- a successful ping to a local address (PC1Bugs’ default gateway – 10.1.1.251)
- how to check the arp cache using the arp command
- a successful ping to a remote address (PC2Sam – 10.1.2.1)
- command abbreviation
It helps if you understand that the reason that your VPCs have any
configuration is because there is a startup file that serves as a script
file when you run vpcs. You will learn more about script fiels in
Lesson #7, but for now, I want you to understand that your startup.vpc
script ended with the show command, and that you can use the show
command any time you want to check your configuraiton, so begin by
practicing that command.
VPCS[1]> show
NAME IP/CIDR GATEWAY MAC LPORT RPORT
VPCS1 10.1.1.1/24 10.1.1.251 00:50:79:66:68:00 20000 30000
fe80::2050:79ff:fe66:6800/64
VPCS2 10.1.2.1/24 10.1.2.252 00:50:79:66:68:01 20001 30001
fe80::2050:79ff:fe66:6801/64
VPCS3 10.1.3.1/24 10.1.3.253 00:50:79:66:68:02 20002 30002
fe80::2050:79ff:fe66:6802/64
VPCS4 0.0.0.0/0 0.0.0.0 00:50:79:66:68:03 20003 30003
fe80::2050:79ff:fe66:6803/64
VPCS5 0.0.0.0/0 0.0.0.0 00:50:79:66:68:04 20004 30004
fe80::2050:79ff:fe66:6804/64
VPCS6 0.0.0.0/0 0.0.0.0 00:50:79:66:68:05 20005 30005
fe80::2050:79ff:fe66:6805/64
VPCS7 0.0.0.0/0 0.0.0.0 00:50:79:66:68:06 20006 30006
fe80::2050:79ff:fe66:6806/64
VPCS8 0.0.0.0/0 0.0.0.0 00:50:79:66:68:07 20007 30007
fe80::2050:79ff:fe66:6807/64
VPCS9 0.0.0.0/0 0.0.0.0 00:50:79:66:68:08 20008 30008
fe80::2050:79ff:fe66:6808/64
Now continue by pinging the default gateway (
10.1.1.251), check the arp cache using the
arp command, then pinging PC2Sam, (VPCS2) whose IP address you can see from above is
10.1.2.1
VPCS[1]> ping 10.1.1.251
10.1.1.251 icmp_seq=1 ttl=255 time=8.741 ms
10.1.1.251 icmp_seq=2 ttl=255 time=3.502 ms
10.1.1.251 icmp_seq=3 ttl=255 time=1.943 ms
10.1.1.251 icmp_seq=4 ttl=255 time=3.289 ms
10.1.1.251 icmp_seq=5 ttl=255 time=2.909 ms
That worked, so there should be an arp entry for 10.1.1.251.
VPCS[1]> arp
c2:00:10:03:00:00 10.1.1.251
The arp entry is OK. When you ping a remote device, often the first
ping times out if a router along the way has go through an arp request.
VPCS[1]> p 10.1.2.1
10.1.2.1 icmp_seq=1 timeout
10.1.2.1 icmp_seq=2 ttl=62 time=8.220 ms
10.1.2.1 icmp_seq=3 ttl=62 time=5.116 ms
10.1.2.1 icmp_seq=4 ttl=62 time=5.171 ms
10.1.2.1 icmp_seq=5 ttl=62 time=6.130 ms
Tip
Notice that in the second ping, I didn’t type the whole word
ping, I abbreviated “
ping” to “
p“. This can be done with any command, so long as you type enough to identify the command.
Lesson #2 – Command history, basic navigation and help
This lesson takes you through:
- The use of the up, down, left & right arrow keys
- The help key (?)
- The hist (history) command
- Changing from one VPC to another
- Aborting output by pressing CTRL+c
If you press the up-arrow then down-arrow keys, you will see that you
can recall your previous commands. You can further edit these commands
using the left-arrow and right-arrow keys. The command
? will give you a page of help, and from that help you can see there is a command
hist which shows a list of the last 50 commands you have used – and hist can be abbreviated to
h.
VPCS[1]> ?
show Print the net configuration of PCs
d Switch to the PC[d], d is digit, range 1 to 9
history List the command history
ip [arguments] Configure PC's IP settings
dhcp Configure host/gateway address using DHCP
arp Show arp table
ping address [options] Ping the network host
tracert address [maxhops] Print the route packets take to network host
echo [text] Display text in output
clear [arguments] Clear ip/ipv6, arp/neighbor cache
set [arguments] Set hostname, connection port and echo on or off
load filename Load the configuration/script from the file 'filename'
save filename Save the configuration to the file 'filename'
ver Show version
? Print help
quit Quit program
VPCS[1]> h
1 show
2 ping 10.1.1.251
3 arp
4 p 10.1.2.1
5 ?
6 h
Tip
Your history is kept from session to session. If you quit a VPCs
session, it saves the current command history in a file called
vpcs.hist, so even when you run this lab next time, your command history
will be preserved from last time!
To access one of the other VPCs, type a digit on a line by itself. In
the example below, notice how I enter 2 to move to PC2, and then use
the up-arrow to retrieve the ping command.
VPCS[1]> 2
VPCS[2]> ping 10.1.1.251
10.1.1.251 icmp_seq=1 ttl=254 time=4.113 ms
10.1.1.251 icmp_seq=2 ttl=254 time=7.437 ms
^C
Tip
You can stop a ping (or tracert) command by pressing
CTRL+c. Notice how the last ping shows only two ping replies, then
^C on the last line, indicating that the ping was interrupted
Lesson #3 – Unsuccessful pings
This lesson will illustrate:
- An arp request failure trying to reach a non-existant local address
- A ping timeout trying to reach a non-existant remote address
- An ICMP type:3, code:1 (Destination Host Unreachable) reply
This time you will ping three non-existant addresses, one on PC1Bugs’
local subnet, one to another subnet off the Yosemite router in the
scenario, and the other to an address that is not part of this network
at all.
VPCS[2]> 1
VPCS[1]> ping 10.1.1.2
host (10.1.1.2) not reachable
VPCS[1]> ping 10.1.2.2
10.1.2.2 icmp_seq=1 timeout
10.1.2.2 icmp_seq=2 timeout
10.1.2.2 icmp_seq=3 timeout
10.1.2.2 icmp_seq=4 timeout
10.1.2.2 icmp_seq=5 timeout
VPCS[1]> ping 4.4.4.4
*10.1.1.251 icmp_seq=1 ttl=255 time=4.660 ms (ICMP type:3, code:1, Destination host unreachable)
*10.1.1.251 icmp_seq=2 ttl=255 time=2.389 ms (ICMP type:3, code:1, Destination host unreachable)
*10.1.1.251 icmp_seq=3 ttl=255 time=2.793 ms (ICMP type:3, code:1, Destination host unreachable)
*10.1.1.251 icmp_seq=4 ttl=255 time=3.353 ms (ICMP type:3, code:1, Destination host unreachable)
*10.1.1.251 icmp_seq=5 ttl=255 time=3.166 ms (ICMP type:3, code:1, Destination host unreachable)
Notice that the first ping says that the local IP (10.1.1.2) was not
reachable – in other words, VPCs sent an ARP request, but did not
receive a reply. See how this is different to the second ping, which
would have been sent all the way to Yosemite router, and of course
Yosemite would not have been able to reach the non-existant 10.1.2.2, so
the pings timeout.
For the third ping to 4.4.4.4, the ICMP packet was sent to the default gateway as well, but this time the
gateway
did not have a path to 4.4..4.4, so the gateway sent back an ICMP
message, type:3, code:1, which equates to Destination Unreachable
(type=3), Host Unreachable (code=1). (See
http://www.iana.org/assignments/icmp-parameters
for a list of ICMP code/type numbers)
Configuration Update
To explore the many other ICMP replies you might get, you will first
have to change some of the routing tables on these routers to make them
think the network is different to what it physically is. In summary,
you will:
- Tell router Albuquerque that the 2.0.0.0/8 network is reachable via
the Seville router (so the Seville router will send back Destination
Unreachable messages)
- Tell router Albuquerque that the 3.0.0.0/8 network is reachable via
the Seville router, AND tell the Seville router that 3.0.0.0/8 is
reachable via Albuquerque (setting up a routing loop, so packets sent to
a 3.x.x.x address will loop until the TTL expires)
- Apply an access list on Seville to stop any TCP/UDP packets on port
80 reaching PC3Elmer (so the Seville router will send ICMP Destination
Administratively Prohibited messages back to the sender)
- Make sure Albuquerque is NOT an http server, so when you send a TCP ping to its port 80, it will reply with a TCP RST
- Make sure Albuquerque is both a tcp small server, so when you send a TCP ping to the TCP echo port (port 7) it will reply
- Make sure Seville is an http server, so when you send a TCP ping to its port 80, it will reply
To do this, cut and paste the following lines into the configuration of Albuquerque router:
enable
configure terminal
ip route 2.0.0.0 255.0.0.0 10.1.130.253
ip route 3.0.0.0 255.0.0.0 10.1.130.253
no ip http server
service tcp-small-servers
end
And cut and paste the following lines into the configuration of Seville router:
enable
configure terminal
ip route 3.0.0.0 255.0.0.0 10.1.130.251
access-list 101 deny tcp any host 10.1.3.1 eq 80
access-list 101 deny udp any host 10.1.3.1 eq 80
access-list 101 permit ip any any
interface fa0/0
ip access-group 101 out
exit
ip http server
end
Lesson #4 – Explore ICMP replies and options
This lesson shows you how to interpret a variety of ICMP replies and options. In particular you will explore:
- an ICMP type:3, code:1 (Destination Host Unreachable) reply from a remote router
- an ICMP type:11, code:0 (TTL expired) reply
- viewing the options of the ping command
- controlling the TTL of the ping packets you send
To achieve this you will need to:
- send a ping from PC1Bugs to something on the 2.0.0.0 network.
Albuquerque should now send it on to Seville, and Seville reply with an
ICMP Destination Unreachable.
- send a ping from PC1Bugs to something on the 3.0.0.0 network.
Albuquerque should now send it on to Seville, and Seville return it
to Albuquerque and so on.
- issue the ping command without an ip address to see the help about the ping options.
- issue the ping command with the -T option to control the TTL values.
Start with a ping to something on the 2.0.0.0 network.
VPCS[1]> ping 2.2.2.2
**10.1.130.253 icmp_seq=1 ttl=254 time=5.326 ms (ICMP type:3, code:1, Destination host unreachable)
*10.1.130.253 icmp_seq=2 ttl=254 time=9.323 ms (ICMP type:3, code:1, Destination host unreachable)
*10.1.130.253 icmp_seq=3 ttl=254 time=6.267 ms (ICMP type:3, code:1, Destination host unreachable)
*10.1.130.253 icmp_seq=4 ttl=254 time=5.789 ms (ICMP type:3, code:1, Destination host unreachable)
Note how the pings went to Seville, and Seville (
10.1.130.253) replied with the ICMP Destination Unreachable messages (type:3, code:1) as expected.
Now test the loop condition you set up.
VPCS[1]> ping 3.3.3.3
*10.1.130.253 icmp_seq=1 ttl=254 time=83.237 ms (ICMP type:11, code:0, TTL expired in transit)
*10.1.130.253 icmp_seq=2 ttl=254 time=60.351 ms (ICMP type:11, code:0, TTL expired in transit)
*10.1.130.253 icmp_seq=3 ttl=254 time=54.258 ms (ICMP type:11, code:0, TTL expired in transit)
*10.1.130.253 icmp_seq=4 ttl=254 time=72.232 ms (ICMP type:11, code:0, TTL expired in transit)
*10.1.130.253 icmp_seq=5 ttl=254 time=74.451 ms (ICMP type:11, code:0, TTL expired in transit)
Note how the pings looped until the TTL expired at Seville, which
sent back ICMP TTL Expired messages (type:11). Try that again, but this
time change the TTL of the packets sent so that they start with an odd
number rather than the default even number (64). Issue a
ping command by itself to see the ping options.
VPCS[1]> ping
ping address [options], Ping the network host, Ctrl+C to stop the command
-1 ICMP mode, default
-2 UDP mode
-3 TCP mode
-P [protocol] Same as above, setting ip protocol
1 - icmp, 17 - udp, 6 - tcp
-c count packet count
-l size data size
-T ttl set TTL, default 64
-p port source and destination port
-f flag tcp head flag, |C|E|U|A|P|R|S|F|
bits |7 6 5 4 3 2 1 0|
-t send packet until interrupt by Ctrl+C
-i ms wait 'ms' milliseconds between sending each packet
-w ms wait 'ms' milliseconds to receive the response
Notice that the option to control the starting TTL value is -T. Use it in the following command.
VPCS[1]> ping 3.3.3.3 -T 3
*10.1.130.251 icmp_seq=1 ttl=255 time=6.718 ms (ICMP type:11, code:0)
*10.1.130.251 icmp_seq=2 ttl=255 time=4.172 ms (ICMP type:11, code:0)
^C
Note that this time, the TTL Expired (Type=11) message came
from Albuquerque (10.1.130.252), because the TTL expired after 3 hops.
Lesson #5 – TCP & UDP pinging
This lesson shows you how to send packets to TCP and UDP ports, and interpret the results. You will:
- use the -P and -p options with the ping command
- see a successful VPCs TCP ping connection
- learn how VPCs TCP ping works
- observe an ICMP Administratively Denied (type:3, code:13) reply to a TCP ping
- observe a TCP RST reply
- see a successful VPCs UDP ping
- observe that an unsuccessful UDP ping draws a ICMP Destination Port Unreachable (type:3, code:3)
To test the Access Control List (ACL), you will use one of VPCs most
powerful features – the ability to send packets to TCP and UDP ports.
To do this you will:
- add a -P 6 (protocol=6, TCP) and a -p 23 (TCP port 23, telnet) and a -p 80 (TCP port 80, HTTP) to your “TCP” pings to see what happens when a TCP “ping” gets through, and when a TCP is denied by an ACL.
- test a TCP “ping” to the Seville router as well, because you made it a HTTP server when you pasted the ip http server command.
- see what happens when you try to send a “TCP ping” to Albuquerque
TCP port 80 (Albuquerque is NOT an HTTP server – you made sure by
pasting a no ip http server command)
- observe the result when you send a TCP ping to the default TCP port
(which happens to be port 7) on Albuquerque – that should work because
you issued a service tcp-small-servers command on Albuquerque.
So here’s the plan:
From PC1Bugs
- Ping PC3Elmer’s TCP port 23 – that should work
- Ping PC3Elmer’s TCP port 80 – that should be denied by the ACL
- Ping Seville router’s TCP port 80 – that should work, because you issued an ip http server command.
- Ping Albuquerque router’s TCP port 80 – that should draw a TCP reset, because issued a no ip http server command. It is NOT an http server.
- Ping Albuquerque router’s default TCP port (TCP port 7 is the TCP echo port) – that should work, because you issued a service tcp-small-servers command.
VPCS[1]> ping 10.1.3.1 -P 6 -p 23
Connect 23@10.1.3.1 seq=1 ttl=62 time=1311.560 ms
SendData 23@10.1.3.1 seq=1 ttl=62 time=12.115 ms
Close 23@10.1.3.1 seq=1 ttl=62 time=10.617 ms
Connect 23@10.1.3.1 seq=2 ttl=62 time=9.944 ms
SendData 23@10.1.3.1 seq=2 ttl=62 time=10.744 ms
Close 23@10.1.3.1 seq=2 ttl=62 time=17.752 ms
Connect 23@10.1.3.1 seq=3 ttl=62 time=8.096 ms
SendData 23@10.1.3.1 seq=3 ttl=62 time=12.436 ms
Close 23@10.1.3.1 seq=3 ttl=62 time=15.322 ms
Connect 23@10.1.3.1 seq=4 ttl=62 time=12.134 ms
SendData 23@10.1.3.1 seq=4 ttl=62 time=20.260 ms
Close 23@10.1.3.1 seq=4 ttl=62 time=17.187 ms
Connect 23@10.1.3.1 seq=5 ttl=62 time=13.777 ms
SendData 23@10.1.3.1 seq=5 ttl=62 time=11.590 ms
Close 23@10.1.3.1 seq=5 ttl=62 time=14.084 ms
A VPCs TCP ping works like this:
- A TCP SYN is sent, and if a TCP SYN/ACK is received, the VPC finishes the connection with an ACK and displays Connect, along with the time taken.
- A data packet (containing a few CR characters) is sent, and if TCP ACK is received, SendData is displayed, along with the time taken.
- A TCP FIN/ACK is sent, and if both an ACK and FIN/ACK are returned,
the VPC finishes the termination handshake with an ACK and displays Close, along with the time taken.
Sometimes the remote system, if it is not another VPC, may respond with a
RST rather than a FIN/ACK, in which case VPC will display Close xx@x.x.x.x timeout
VPCS[1]> ping 10.1.3.1 -P 6 -p 80
*10.1.130.253 tcp_seq=1 ttl=254 time=9.455 ms (ICMP type:3, code:13, Communication administratively prohibited)
*10.1.130.253 tcp_seq=3 ttl=254 time=3.924 ms (ICMP type:3, code:13, Communication administratively prohibited)
*10.1.130.253 tcp_seq=5 ttl=254 time=6.600 ms (ICMP type:3, code:13, Communication administratively prohibited)
Note how the Seville router sent an ICMP Destination Unreachable
(type:3) Administratively Prohibited (code:13) message for the TCP ping
to PC3Elmer’s port 80. What a great tool to test ACLs!!
VPCS[1]> ping 10.1.3.253 -P 6 -p 80
Connect 80@10.1.3.253 seq=1 ttl=254 time=12.448 ms
SendData 80@10.1.3.253 seq=1 ttl=254 time=215.492 ms
Close 80@10.1.3.253 seq=1 ttl=254 time=16.636 ms
Connect 80@10.1.3.253 seq=2 ttl=254 time=5.434 ms
SendData 80@10.1.3.253 seq=2 ttl=254 time=234.653 ms
Close 80@10.1.3.253 seq=2 ttl=254 time=13.759 ms
Connect 80@10.1.3.253 seq=3 ttl=254 time=6.927 ms
SendData 80@10.1.3.253 seq=3 ttl=254 time=232.419 ms
Close 80@10.1.3.253 seq=3 ttl=254 time=10.550 ms
Connect 80@10.1.3.253 seq=4 ttl=254 time=6.130 ms
SendData 80@10.1.3.253 seq=4 ttl=254 time=235.334 ms
Close 80@10.1.3.253 seq=4 ttl=254 time=11.846 ms
Connect 80@10.1.3.253 seq=5 ttl=254 time=5.188 ms
SendData 80@10.1.3.253 seq=5 ttl=254 time=202.523 ms
Close 80@10.1.3.253 seq=5 ttl=254 time=12.864 ms
That proves that Seville is indeed listening on port 80!
VPCS[1]> ping 10.1.1.251 -P 6 -p 80
Connect 80@10.1.1.251 RST returned
Connect 80@10.1.1.251 RST returned
Connect 80@10.1.1.251 RST returned
Connect 80@10.1.1.251 RST returned
Connect 80@10.1.1.251 RST returned
As expected, Albuquerque is NOT listening on port 80, and says so!
VPCS[1]> ping 10.1.1.251 -P 6
Connect 7@10.1.1.251 seq=1 ttl=255 time=8.001 ms
SendData 7@10.1.1.251 seq=1 ttl=255 time=6.384 ms
Close 7@10.1.1.251 seq=1 ttl=255 time=9.536 ms
Connect 7@10.1.1.251 seq=2 ttl=255 time=4.909 ms
SendData 7@10.1.1.251 seq=2 ttl=255 time=4.741 ms
Close 7@10.1.1.251 seq=2 ttl=255 time=6.241 ms
Connect 7@10.1.1.251 seq=3 ttl=255 time=3.352 ms
SendData 7@10.1.1.251 seq=3 ttl=255 time=4.962 ms
Close 7@10.1.1.251 seq=3 ttl=255 time=9.370 ms
Connect 7@10.1.1.251 seq=4 ttl=255 time=4.476 ms
SendData 7@10.1.1.251 seq=4 ttl=255 time=5.269 ms
Close 7@10.1.1.251 seq=4 ttl=255 time=7.219 ms
Connect 7@10.1.1.251 seq=5 ttl=255 time=4.735 ms
SendData 7@10.1.1.251 seq=5 ttl=255 time=7.637 ms
Close 7@10.1.1.251 seq=5 ttl=255 time=7.043 ms
Note that the last ping didn’t specify a port number – I just
specified the protocol, (-P 6) and the VPC sent it to TCP port 7, which
is the well-known TCP echo port.
Tip
VPCs has some shortcuts for IP protocols 1, 6, & 17 – see the output of the ping help above. I could have said
ping 10.1.1.251 -3 instead of
ping 10.1.1.251 -P 6.
Personally I prefer -P 6, because it makes me remember that TCP is IP
protocol #6, which may be handy information in an exam one day!
And as you can see from the output above, all predictions were correct. That is:
- The Ping to PC3Elmer’s TCP port 23 worked
- The Ping to PC3Elmer’s TCP port 80 was denied by the ACL
- The Ping to Seville router’s TCP port 80 worked.
- The Ping to Albuquerque router’s TCP port 80 returned a TCP reset, because it is NOT an http server.
- The Ping to Albuquerque router’s default TCP port worked, because you issued a service tcp-small-servers command.
Now let’s try some UDP pings. This time you will see some new ICMP
replies – specifically ICMP Destination Port unreachable. When you send
a TCP SYN packet to a device that is not listening on a particular
port, the target device sends back a TCP RST (reset) segment. However,
if you send a UDP packet to a device that is not listening on a
particular port, the target device sends back am ICMP type:3, code:3 –
Destination Port unreachable.
So here’s the plan:
From PC1Bugs
- Ping PC3Elmer’s UDP port 99 – that should work – VPCs respond to all UDP packets from other VPCs
- Ping PC3Elmer’s UDP port 80 – that should be denied by the ACL
- Ping Seville router’s UDP port 99 – that should see an ICMP Destination Port Unreachable reply (type:3, code:3)
VPCS[1]> ping 10.1.3.1 -P 17 -p 99
10.1.3.1 udp_seq=1 ttl=62 time=7.920 ms
10.1.3.1 udp_seq=2 ttl=62 time=5.623 ms
10.1.3.1 udp_seq=3 ttl=62 time=4.979 ms
10.1.3.1 udp_seq=4 ttl=62 time=8.003 ms
10.1.3.1 udp_seq=5 ttl=62 time=6.732 ms
VPCS[1]> ping 10.1.3.1 -P 17 -p 80
**10.1.130.253 udp_seq=1 ttl=254 time=6.053 ms (ICMP type:3, code:13, Communication administratively prohibited)
*10.1.130.253 udp_seq=2 ttl=254 time=4.337 ms (ICMP type:3, code:13, Communication administratively prohibited)
^C
VPCS[1]> ping 10.1.3.253 -P 17 -p 99
*10.1.3.253 udp_seq=1 ttl=254 time=12.187 ms (ICMP type:3, code:3, Destination port unreachable)
*10.1.3.253 udp_seq=2 ttl=254 time=6.813 ms (ICMP type:3, code:3, Destination port unreachable)
^C
And as you can see from the output, all predictions were correct. That is:
- The Ping to PC3Elmer’s UDP port 99 worked
- The Ping to PC3Elmer’s UDP port 80 was denied by the ACL
- The Ping to Seville router’s TCP port 99 caused an ICMP Destination Port Unreachable reply (type:3, code:3) reply.
Lesson #6 – Trying Tracert (traceroute)
Tracert actually uses the fact that routers send ICMP destination
unreachable messages to trace the path through a network. In this
lesson, you will:
- watch a tracert succeed
- watch a tracert strike a ICMP Destination Unreachable reply along the path
- watch a tracert detect a loop
- control the number of hops that tracert will trace for
Using the existing topology, trace the path from PC1Bugs to
- PC2Sam (10.1.2.1) – this should succeed
- unknown remote address, 4.4.4.4 and 2.2.2.2 – you can expect that
PC1Bugs’ default gateway (Albuquerque) will reply with a Destination
Unreachable for the 4.4.4.4 address, and since you added a route
to Albuquerque directing traffic to the 2.0.0.0 network to Seville, you
can expect that Seville will reply with a Destination Unreachable for
the trace to 2.2.2.2
- unknown remote address, 3.3.3.3 – you can expect that Albuquerque
will forward this to Seville, and Seville pass it back to Albuquerque
and so on because I have engineered the loop with the route statements
you added earlier.
VPCS[1]> trace 4.4.4.4
traceroute to 4.4.4.4, 64 hops max, press Ctrl+C to stop
1 10.1.1.251 3.669 ms 1.977 ms 3.064 ms
2 *10.1.1.251 2.238 ms (ICMP type:3, code:1, Destination host unreachable)
The ICMP type:3, code:1 is an ICMP Destination Host Unreachable, as was predicted.
Note that the default gateway router (Albuquerque) did actually reply to
the first round of packets sent with a TTL of 1, proving that routers
actually make the routing decision
before they decrement the TTL.
Note that reply 2 has a * character at the beginning of the line to indicate that this reply indicates a failure of some kind.
VPCS[1]> trace 2.2.2.2
traceroute to 2.2.2.2, 64 hops max, press Ctrl+C to stop
1 10.1.1.251 2.579 ms 3.204 ms 3.462 ms
2 10.1.130.253 7.917 ms 4.209 ms 4.673 ms
3 *10.1.130.253 12.890 ms (ICMP type:3, code:1, Destination host unreachable)
This time the trace got to Seville, but
Seville has no route to the 2.0.0.0 network, so replied with the ICMP
Destination Host Unreachable (type:3, code:1)
VPCS[1]> trace 3.3.3.3
traceroute to 3.3.3.3, 64 hops max, press Ctrl+C to stop
1 10.1.1.251 2.979 ms 1.700 ms 2.143 ms
2 10.1.130.253 6.280 ms 4.980 ms 5.108 ms
3 10.1.130.251 9.438 ms 9.512 ms 4.398 ms
4 10.1.130.253 4.758 ms 6.241 ms 6.453 ms
5 10.1.130.251 6.460 ms 10.098 ms 7.811 ms
6 10.1.130.253 8.771 ms 9.053 ms 8.533 ms
7 10.1.130.251 11.545 ms 13.584 ms 9.932 ms
8 10.1.130.253 10.501 ms 11.806 ms 12.095 ms
9 10.1.130.251 55.977 ms 60.101 ms^C *
As expected, the trace ran back and forth between Albuquerque and Seville.
Note I had to stop the trace by hitting CTRL+c, but I could have limited the number of replies by specifying the
maxhops option.
VPCS[1]> trace 3.3.3.3 4
traceroute to 3.3.3.3, 4 hops max, press Ctrl+C to stop
1 10.1.1.251 3.577 ms 2.185 ms 1.404 ms
2 10.1.130.253 5.021 ms 3.862 ms 5.262 ms
3 10.1.130.251 4.358 ms 4.557 ms 4.890 ms
4 10.1.130.253 9.543 ms 9.636 ms 5.829 ms
By specifying the number 4 at the end of the
tracert command, I didn’t need to hit CTRL+c to stop the trace this time.
Lesson #7 – Fun Stuff
There are a few other features that haven’t been explored yet, but
can be very useful especially for documentation and test scripts. You
are about to explore:
- Changing the VPCs display name
- Changing a VPCs IP address
- Creating a VPCs script file
- Running a VPCs script file
- Editing a VPCs script
- How to use the echo command
- How to control the output of scripts the set echo on and set echo off commands
In this lesson you will:
- Change VPCs 1 -3 to names to match the diagram – ie PC1=Bugs, PC2=Sam and PC3=Elmer using the set pcname command.
- Change the IP address of PC2Sam to 10.1.2.2/24
- Save your configuration to a file called script1 using the save command
- Execute a series of commands to load them into your history buffer
- Quit VPCs to save your history buffer
- Use a text editor to combine parts of your history buffer (vpcs.hist) with your script file (script1) to create a test script called script2
- Document your script with comments and echo commands
- load the text script2 into VPCs using the load command
- change the behaviour of your script by using the set echo on and set eco off commands
Start with the name changing as shown below. Note that you can get help about the set command by typing
set, and that the first attempt to set the pcname to
PC1Bugs shows you that the maximum length of a pcname is 6 characters
VPCS[1]> set
set [lport|rport|pcname|echo], Set connection port, hostname or echo setting
lport port local port, listen by VPCS
rport port remote port, listen by dynamips
pcname name rename the current pc
echo [on|off] set echoing on or off
VPCS[1]> set pcname PC1Bugs
Hostname is too long. (should be less than 6)
VPCS[1]> set pcname Bugs
Bugs[1]> 2
VPCS[2]> set pcname Sam
Sam[2]> 3
VPCS[3]> set pcname Elmer
Elmer[3]>
So far, the PC IP addresses have stayed constant, so for practice, change Sam’s IP address to
10.1.2.2, leave the gateway as 10.1.2.252 and set the mask to 24 bits
Elmer[3]> 2
Sam[2]> ip
ip [dhcp|auto|address], Configure PC's IP settings
dhcp Configure host/gateway address using DHCP, only ipv4
auto Stateless address autoconfiguration, only ipv6
PC will try to get the ipv6 address from the router at startup
address [gateway] [CIDR]
set the PC's ip, gateway's ip and network mask
Default IPv4 CIDR is 24, IPv6 is 64. In the ether mode,
the ip of the tapx is the maximum host ID of the subnet.
'ip 10.1.1.70 10.1.1.65 26', set the host ip to 10.1.1.70,
the gateway ip to 10.1.1.65, the netmask to 255.255.255.192,
the tapx ip to 10.1.1.126 in the ether mode.
Sam[2]> ip 10.1.2.2 10.1.2.252 24
PC2 : 10.1.2.2 255.255.255.0 gateway 10.1.2.252
Note that since the default mask is 24 bits, I could have just issued the command,
ip 10.1.2.2 10.1.2.152
Now that you have changed the hostnames and one of the IP addresses,
it is a good time to save your configuration, and quit VPCs, which will
force VPCs to save the history file, which you will need in the next
step.
Sam[2]> save script1
......... done
Sam[2]> quit
Of course, if you are running VPCs from within the GNS3 Workbench
environment, VPCs will actually automatically restart, because I have
run this instance of VPCs from within a script which saves people from
accidentally quitting. However that doesn’t matter, because you saved
your configuration remember?
So now reload your configuration, and check it with the show command.
You should see all your pcnames restored, and the ip address for Sam
has changed.
VPCS[1]> load script1
Executing the startup file
PC1 : 10.1.1.1 255.255.255.0 gateway 10.1.1.251
PC2 : 10.1.2.2 255.255.255.0 gateway 10.1.2.252
PC3 : 10.1.3.1 255.255.255.0 gateway 10.1.3.253
VPCS[9]> showNAME IP/CIDR GATEWAY MAC LPORT RPORT
Bugs 10.1.1.1/24 10.1.1.251 00:50:79:66:68:00 20000 30000
fe80::2050:79ff:fe66:6800/64
Sam 10.1.2.2/24 10.1.2.252 00:50:79:66:68:01 20001 30001
fe80::2050:79ff:fe66:6801/64
Elmer 10.1.3.1/24 10.1.3.253 00:50:79:66:68:02 20002 30002
fe80::2050:79ff:fe66:6802/64
VPCS4 0.0.0.0/0 0.0.0.0 00:50:79:66:68:03 20003 30003
fe80::2050:79ff:fe66:6803/64
VPCS5 0.0.0.0/0 0.0.0.0 00:50:79:66:68:04 20004 30004
fe80::2050:79ff:fe66:6804/64
VPCS6 0.0.0.0/0 0.0.0.0 00:50:79:66:68:05 20005 30005
fe80::2050:79ff:fe66:6805/64
VPCS7 0.0.0.0/0 0.0.0.0 00:50:79:66:68:06 20006 30006
fe80::2050:79ff:fe66:6806/64
VPCS8 0.0.0.0/0 0.0.0.0 00:50:79:66:68:07 20007 30007
fe80::2050:79ff:fe66:6807/64
VPCS9 0.0.0.0/0 0.0.0.0 00:50:79:66:68:08 20008 30008
fe80::2050:79ff:fe66:6808/64
VPCS[9]> 2
Sam[2]>
The next task is to create a new file which will become our second script file, and you will use the contents of the
script1 you saved in VPCs and the contents of
vpcs.hist to create this script. In a file browser, navigate to the folder
/opt/GNS3/Project/3Router_ICND_Book_Example – it is the same folder where you fond the
startup.2 (eigrp configured) file when you started in Lesson #1.
Create a blank file called
script2 in this folder – you can do this by clicking
File->Create Document -> Empty file, and giving it the name
script2.
Double click on this file to open it in gedit. Put a comment right at the top something like this:
#Script file created for testing vpcs
Tip
Comments can be placed in script files by starting the comment with the # character.
Now locate the file called
script1 in the same directory and double-click on it to open it also in gedit. It should look like this:
1
set pcname Bugs
ip 10.1.1.1 10.1.1.251 24
2
set pcname Sam
ip 10.1.2.2 10.1.2.252 24
3
set pcname Elmer
ip 10.1.3.1 10.1.3.253 24
4
set pcname VPCS
5
set pcname VPCS
6
set pcname VPCS
7
set pcname VPCS
8
set pcname VPCS
9
set pcname VPCS
Copy and paste the first 9 lines into your script2 file. It should now look like:
#Script file created for testing vpcs
1
set pcname Bugs
ip 10.1.1.1 10.1.1.251 24
2
set pcname Sam
ip 10.1.2.2 10.1.2.252 24
3
set pcname Elmer
ip 10.1.3.1 10.1.3.253 24
Now open (in the same directory) the file called vpcs.hist. In it
you will find all the commands you have entered so far. If you have
followed this tutorial to the letter, it should look like this:
show
ping 10.1.1.251
arp
p 10.1.2.1
?
h
2
p 10.1.2.1
1
ping 10.1.1.2
ping 10.1.2.2
ping 4.4.4.4
ping 2.2.2.2
ping 3.3.3.3
ping
ping 3.3.3.3 -T 3
ping 10.1.3.1 -P 6 -p 23
ping 10.1.3.1 -P 6 -p 80
ping 10.1.3.253 -P 6 -p 80
ping 10.1.1.251 -P 6 -p 80
ping 10.1.3.1 -P 17 -p 99
ping 10.1.3.1 -P 17 -p 80
ping 10.1.3.253 -P 17 -p 99
trace 4.4.4.4
trace 2.2.2.2
trace 3.3.3.3
trace 3.3.3.3 4
set
set pcname Bugs
2
set pcname Sam
3
set pcname Elmer
2
ip
ip 10.1.2.2 10.1.2.252 24
save script1
quit
Now I want you to copy all of the lines from vpcs.hist EXCEPT the
last 2 lines – the save command and the quit command at the end. You are about to paste them in script2, but there is one more task before we do that!
Back in the script2 file, go to the end of the file and add these commands:
1
echo Here is the beginning of the history file
The command 1 on a line by itself is to ensure that we start at the
correct PC. The echo command will display a message as our script is
executing. Now, immediately after these lines, paste the contents of
vpcs.hist that you just copied, and save the file. Don’t close it, you
have more to do yet.
And now let’s see this script run! Back in VPCs, load this updated script.
load script2
And watch all those commands being executed – did you see the message “Here is the beginning of the history file” appear?
But if you look closely, you will see there is a problem – you can’t see
exactly which commands are producing which output. Never fear there is
an answer for that.
Go back to your script2 file in the editor, and add the single command at the top of the file (after your comment)
set echo on
Save the file again, and load it one more time. If you get sick of waiting for it to complete, hit CTRL+c
Tip
Script files can be aborted by hitting CTRL+c
Now notice that while the script is executing, the command that
produces the output is displayed before the command. This is the effect
of the
set echo on command. However, there is a catch.
Enter a command like
ping 10.1.1.251
Bugs[1]> ping 10.1.1.251
ping 10.1.1.251
10.1.1.251 icmp_seq=1 ttl=255 time=2.756 ms
10.1.1.251 icmp_seq=2 ttl=255 time=2.895 ms
10.1.1.251 icmp_seq=3 ttl=255 time=3.437 ms
10.1.1.251 icmp_seq=4 ttl=255 time=3.104 ms
10.1.1.251 icmp_seq=5 ttl=255 time=4.335 ms
See how the command is echoed on the output after you type it? This can be corrected by reversing the
set echo on command we put at the beginning of the file by putting a
set echo off
at the end of the file – but of course you will have to let the whole
script run without interruption for it to take effect. Of course you
could just type:
set echo off at the vpcs command line to turn it off yourself.
Footnote: What VPCs DOESN’T do
VPCs is a VERY useful troubleshooting tool when used in your GNS3
environment, and does everything you’d want it to do almost 100% of the
time. There are some minor features not implemented that you might want
to be aware of:
- VPCs has no concept of MTU or IP fragmentation. If you ask it to
send a ping of 2000 bytes, it will. All in one packet! This makes it
impossible to use to test IP fragmentation.
- Still on the fragmentation, if you send a VPCs host a ping that
arrives in fragments, it can’t put the fragments back together, so the
pings will fail.
- The DF (Don’t Fragment) bit on your pings is set by default, so you
can test links with MTUs less than 1500 OK – however, if you want to set
fragmentation across that link, you can’t turn the DF bit off.
- Similarly, other fancy IP options, like record route can’t be set either.
- VPCs don’t have any layer 3 routing capability beyond a default
gateway. Therefore VPCs can’t act upon ICMP re-directs if it ever
receives one.
Thanks to http://rednectar.net
