Posts Tagged ‘labs’
Getting BGP routes into dynamips (with video)
Written by jlgaddis on August 21, 2009 – 5:24 am -UPDATE: If you’re not interesting in doing this yourself and just want to see it in action, check out the video, “Dumping 265k BGP routes into dynamips”. When I went through these steps to document, I ended up with 298,870 prefixes in my dynamips router. When I did it the second time, to record the video, I only ended up with 265,857 prefixes. Not sure why the discrepancy, but at least you can see it works! YMMV!
After I posted “Dynamips, a 7200, and a full BGP table”, a number of you left comments asking how I got the BGP routes into dynamips… the answer is einval’s “bgpsimple”.
“This perl script allows to setup an BGP adjacency with a BGP peer, monitor the messages and updates received from that peer, and to send out updates from a predefined set of NLRIs/attributes. BGP session and message handling is done by Net::BGP.”
0. Pre-requisites
On Ubuntu, at least, you’re going to need to install some packages that likely aren’t already installed. We’re going to need these to be able to build bgpdump in step 2. Fortunately, the following command will install everything you need (well, except for Net::BGP and bgpsimple):
[root@stewie ~]# apt-get install build-essential zlib1g-dev libbz2-dev
1. Install Net::BGP
Before we can even think about doing this, we’re going to need to install the Net::BGP perl modules, most likely from CPAN (your distribution may provide it in a handy installable package, but I wouldn’t count on it). I’m using an Ubuntu 8.04 LTS Server installation — you can use whichever distribution (or BSD) that you like, but this is what I’m using.
Fire up the CPAN shell:
[root@stewie ~]# perl -MCPAN -e shell cpan shell -- CPAN exploration and modules installation (v1.9402) Enter 'h' for help. cpan[1]>
If this is the first time you’ve done this, you’ll have to go through some configuration. That configuration is out of scope of this document. Google it.
Next, install Net::BGP and exit the CPAN shell:
cpan[1]> install Net::BGP CPAN: Storable loaded ok (v2.15) Going to read '/home/jlgaddis/.cpan/Metadata' Database was generated on Thu, 20 Aug 2009 22:27:00 GMT Running install for module 'Net::BGP' Running make for K/KB/KBRINT/Net-BGP-0.13.tar.gz [snip] Appending installation info to /usr/lib/perl/5.8/perllocal.pod KBRINT/Net-BGP-0.13.tar.gz /usr/bin/make install -- OK cpan[2]> exit Lockfile removed. [root@stewie ~]#
2. Install bgpdump
As mentioned in bgpsimple’s README, we’re going to use a RIB dump from a router in the default-free zone. Fortunately, RIPE makes this data available for download. Before we can use it, however, we need to convert it to a format that bgpsimple can use. We’re going to download and compile bgpdump which can do the conversion for us.
[root@stewie ~]# wget http://www.ris.ripe.net/source/libbgpdump-1.4.99.9.tar.gz
--22:00:48-- http://www.ris.ripe.net/source/libbgpdump-1.4.99.9.tar.gz
=> `libbgpdump-1.4.99.9.tar.gz'
Resolving www.ris.ripe.net... 193.0.19.19
Connecting to www.ris.ripe.net|193.0.19.19|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 82,909 (81K) [application/x-gzip]
100%[=====================================================>] 82,909 130.32K/s
22:00:49 (129.85 KB/s) - `libbgpdump-1.4.99.9.tar.gz' saved [82909/82909]
[root@stewie ~]#
Uncompress the tarball and change to the newly created directory:
[root@stewie ~]# tar zxf libbgpdump-1.4.99.9.tar.gz [root@stewie ~]# cd libbgpdump-1.4.99.9/
Be sure to skim through the README file in this directory.
Now, we can begin to build bgpdump. I don’t need IPv6 support, so I’m going to leave it out.
[root@stewie ~/libbgpdump-1.4.99.9]# ./configure --disable-ipv6 checking for gcc... gcc checking for C compiler default output file name... a.out checking whether the C compiler works... yes [snip] checking for inet_ntoa... yes checking for inet_ntop... yes checking for IPv6 support... disabled configure: creating ./config.status config.status: creating Makefile config.status: creating bgpdump-config.h [root@stewie ~/libbgpdump-1.4.99.9]#
Once the configure script has completed (successfully!), we can build bgpdump:
[root@stewie ~/libbgpdump-1.4.99.9]# make [snip] [root@stewie ~/libbgpdump-1.4.99.9]# ls -l bgpdump -rwxr-xr-x 1 root root 46540 2009-08-20 22:14 bgpdump
As you see, we end up with a binary named “bgpdump”, which I’m going to copy over to /usr/local/bin. I’m also going to create a directory named “bgp”, where I’ll store the files we’ll be working with:
[root@stewie ~/libbgpdump-1.4.99.9]# cp bgpdump /usr/local/bin [root@stewie ~/libbgpdump-1.4.99.9]# mkdir ../bgp [root@stewie ~/libbgpdump-1.4.99.9]# cd ../bgp
3. Get some route data
Before we can inject any routes into our router, we need some routes to inject! As mentioned, RIPE makes these available to us. Go to the “RIS Raw Data” page, pick a collector, then download a file containing the raw data:
[root@stewie ~/bgp]# wget http://data.ris.ripe.net/rrc16/2009.08/bview.20090820.2359.gz
--22:59:26-- http://data.ris.ripe.net/rrc16/2009.08/bview.20090820.2359.gz
=> `bview.20090820.2359.gz'
Resolving data.ris.ripe.net... 193.0.19.19
Connecting to data.ris.ripe.net|193.0.19.19|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 3,108,638 (3.0M) [application/x-gzip]
100%[=====================================================>] 3,108,638 234.33K/s ETA 00:00
22:59:40 (214.70 KB/s) - `bview.20090820.2359.gz' saved [3108638/3108638]
[root@stewie ~/bgp]#
Now that we have some routing data, we need to get it into a format that bgpsimple can work with. This is where bgpdump comes into play. Copying from bgpsimple’s README:
[root@stewie ~/bgp]# zcat bview.20090820.2359.gz | bgpdump -m - > myroutes [root@stewie ~/bgp]#
4. Download bgpsimple
Download the code for bgpsimple:
[root@stewie ~/bgp]# wget http://bgpsimple.googlecode.com/files/bgp_simple.tgz
--23:11:17-- http://bgpsimple.googlecode.com/files/bgp_simple.tgz
=> `bgp_simple.tgz'
Resolving bgpsimple.googlecode.com... 209.85.225.82
Connecting to bgpsimple.googlecode.com|209.85.225.82|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 9,324 (9.1K) [application/x-gzip]
100%[=====================================================>] 9,324 --.--K/s
23:11:17 (114.18 KB/s) - `bgp_simple.tgz' saved [9324/9324]
[root@stewie ~/bgp]#
Uncompress the tarball:
[root@stewie ~/bgp]# tar zxf bgp_simple.tgz
You should end up with a Perl script named “bgp_simple.pl”:
[root@stewie ~/bgp]# ls -l bgp_simple.pl -rwxr-xr-x 1 jlgaddis jlgaddis 20388 2009-01-07 10:31 bgp_simple.pl
5. Start up your dynamips router
Now it’s time to fire up our virtual 7200 router. Here’s the .net file for dynagen that I used (don’t forget to change the filenames and paths, as appropriate).
Start up dynamips, start up dynagen, connect to the console, and do some initial configuration:
[jlgaddis@stewie ~]$ telnet 192.168.1.109 2000
Trying 192.168.1.109...
Connected to 192.168.1.109.
Escape character is '^]'.
Connected to Dynamips VM "R1" (ID 0, type c7200) - Console port
Restricted Rights Legend
Use, duplication, or disclosure by the Government is
subject to restrictions as set forth in subparagraph
(c) of the Commercial Computer Software - Restricted
Rights clause at FAR sec. 52.227-19 and subparagraph
(c) (1) (ii) of the Rights in Technical Data and Computer
Software clause at DFARS sec. 252.227-7013.
[snip]
Router> enable
Router# configure terminal
Router(config)# no ip domain lookup
Router(config)# no ip http server
Router(config)# hostname c7200
c7200(config)# line con 0
c7200(config-line)# exec-timeout 0 0
c7200(config-line)# logging synchronous
6. Configure dynamips router’s network interface
We need to put an IP address on the router’s fastethernet 2/0 interface, then verify that we can ping the host that we’re going to run bgpsimple on:
c7200(config-line)# interface fastethernet 2/0 c7200(config-if)# ip address 192.168.1.99 255.255.255.0 c7200(config-if)# no shutdown c7200(config-if)# *Aug 20 23:23:26.167: %LINK-3-UPDOWN: Interface FastEthernet2/0, changed state to up *Aug 20 23:23:27.167: %LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet2/0, changed state to up c7200(config-if)# do ping 192.168.1.104 Type escape sequence to abort. Sending 5, 100-byte ICMP Echos to 192.168.1.104, timeout is 2 seconds: !!!!! Success rate is 100 percent (5/5), round-trip min/avg/max = 4/8/12 ms
7. Configure the BGP session on your dynamips router
Next, we need to configure our dynamips router for a BGP session with bgpsimple. You will need to change IP address below (192.168.1.104) to the IP address of the box you are running bgpsimple on:
c7200(config-if)# router bgp 65000 c7200(config-router)# no synchronization c7200(config-router)# no auto-summary c7200(config-router)# neighbor 192.168.1.104 remote-as 65000
8. Test with a limited number of prefixes
Now that our dynamips router is configured for BGP, we’re ready for a quick test with a small number of prefixes (10, for now). Look at the README for what all these command-line options mean (I wrapped this for readability, you don’t have to):
[root@stewie ~/bgp]# ./bgp_simple.pl -myas 65000 -myip 192.168.1.104 \ > -peerip 192.168.1.99 -peeras 65000 -p myroutes -m 10 -n ---------------------------------------- CONFIG SUMMARY -------------------------------------------------- Configured for an iBGP session between me (ASN65000, 192.168.1.104) and peer (ASN65000, 192.168.1.99). Will use prefixes from file myroutes. Maximum number of prefixes to be advertised: 10. Will spoof next hop address to 192.168.1.104. ---------------------------------------------------------------------------------------------------------- Sending full update. [snip]
Looks like that worked, let’s take a look at the BGP table on our dynamips router:
*Aug 20 23:31:49.715: %BGP-5-ADJCHANGE: neighbor 192.168.1.104 Up
c7200# show ip bgp
BGP table version is 31, local router ID is 192.168.1.99
Status codes: s suppressed, d damped, h history, * valid, > best, i - internal,
r RIB-failure, S Stale
Origin codes: i - IGP, e - EGP, ? - incomplete
Network Next Hop Metric LocPrf Weight Path
*>i1.1.1.0/24 192.168.1.104 0 0 25152 6939 3303 8300 i
*>i3.0.0.0 192.168.1.104 0 0 25152 6939 15412 9304 80 i
*>i3.51.92.0/23 192.168.1.104 0 0 25152 23148 7018 ?
*>i4.0.0.0/9 192.168.1.104 0 0 25152 1273 3356 i
*>i4.0.0.0 192.168.1.104 0 0 25152 1273 3356 i
*>i4.21.103.0/24 192.168.1.104 0 0 25152 6939 3549 46133 i
*>i4.23.88.0/24 192.168.1.104 0 0 25152 23148 7018 46164 i
*>i4.23.88.0/23 192.168.1.104 0 0 25152 23148 7018 46164 i
*>i4.23.89.0/24 192.168.1.104 0 0 25152 23148 7018 46164 i
*>i4.23.92.0/22 192.168.1.104 0 0 25152 23148 7018 46164 i
c7200#
And there’s our 10 routes! w00t!
9. Advertise all the routes!
Now that we now we can get an adjacency up and exchange routes, let’s go for the gusto!
Kill bgp_simple.pl (CTRL-C works) and let’s take a quick look at how many routes are in the “myroutes” file.
[root@stewie ~/bgp]# wc -l myroutes 300035 myroutes
In my case, we have just over 300k. Your numbers may vary slightly — and there very well may be duplicate prefixes — depending on which dump you download from RIPE. In order to inject all the routes, we just run bgp_simple.pl as before, but without the “-m 10″ (maximum of 10 prefixes to advertise) option (again, wrapped for readability):
[root@stewie ~/bgp]# ./bgp_simple.pl -myas 65000 -myip 192.168.1.104 \ > -peerip 192.168.1.99 -peeras 65000 -p myroutes -n ---------------------------------------- CONFIG SUMMARY -------------------------------------------------- Configured for an iBGP session between me (ASN65000, 192.168.1.104) and peer (ASN65000, 192.168.1.99). Will use prefixes from file myroutes. Maximum number of prefixes to be advertised: 10. Will spoof next hop address to 192.168.1.104. ---------------------------------------------------------------------------------------------------------- Sending full update. [snip]
And now we just watch the number of prefixes received continually go up on our dynamips router:
c7200# show ip bgp summary | begin Neighbor Neighbor V AS MsgRcvd MsgSent TblVer InQ OutQ Up/Down State/PfxRcd 192.168.1.104 4 65000 98425 21 98413 0 0 00:01:24 98237
c7200# show ip bgp summary | begin Neighbor Neighbor V AS MsgRcvd MsgSent TblVer InQ OutQ Up/Down State/PfxRcd 192.168.1.104 4 65000 141060 23 141069 0 0 00:02:04 140849
c7200# show ip bgp summary | begin Neighbor Neighbor V AS MsgRcvd MsgSent TblVer InQ OutQ Up/Down State/PfxRcd 192.168.1.104 4 65000 238966 25 238975 0 0 00:03:16 238739
We can take a quick look at a few of the prefixes in our BGP table:
c7200# show ip bgp 12.0.0.0/8 longer-prefixes | begin Network Network Next Hop Metric LocPrf Weight Path *>i12.0.0.0/9 192.168.1.104 0 0 25152 23148 7018 i *>i12.0.0.0 192.168.1.104 0 0 25152 23148 7018 i *>i12.0.18.0/24 192.168.1.104 0 0 25152 23148 7018 27585 i *>i12.0.19.0/24 192.168.1.104 0 0 25152 1273 3561 27487 i *>i12.0.28.0/24 192.168.1.104 0 0 25152 1273 4323 30050 i *>i12.0.29.0/24 192.168.1.104 0 0 25152 1273 174 30538 i *>i12.0.33.0/24 192.168.1.104 0 0 25152 1273 174 40544 i *>i12.0.43.0/24 192.168.1.104 0 0 25152 23148 7018 2386 i *>i12.0.48.0/20 192.168.1.104 0 0 25152 1273 174 1742 i *>i12.0.153.0/24 192.168.1.104 0 0 25152 23148 7018 6519 i *>i12.0.170.0/24 192.168.1.104 0 0 25152 23148 7018 22528 i *>i12.0.239.0/24 192.168.1.104 0 0 25152 19151 1239 33628 i [snip]
That’s all there is to it!
10. All your routes are belong to us!
c7200# show ip bgp summary BGP router identifier 192.168.1.99, local AS number 65000 BGP table version is 1038362, main routing table version 1038362 298870 network entries using 34967790 bytes of memory 298870 path entries using 15541240 bytes of memory 51910/51909 BGP path/bestpath attribute entries using 6436840 bytes of memory 47723 BGP AS-PATH entries using 1265376 bytes of memory 1 BGP community entries using 24 bytes of memory 0 BGP route-map cache entries using 0 bytes of memory 0 BGP filter-list cache entries using 0 bytes of memory BGP using 58211270 total bytes of memory BGP activity 668359/369489 prefixes, 668359/369489 paths, scan interval 60 secs Neighbor V AS MsgRcvd MsgSent TblVer InQ OutQ Up/Down State/PfxRcd 192.168.1.104 4 65000 669061 20 1038362 0 0 00:04:16 298870 c7200#
OPTIONAL: If the BGP connection between bgpsimple and your dynamips router dies due to a hold timer expiring, you can get around this by doing two things:
- Editing bgp_simple.pl. At line 220, I added two lines:
KeepAliveTime => 600,
HoldTime => 1800,
That “section” of code (setting up the peer connection using Net::BGP::Peer) now reads like this:
my $bgp = Net::BGP::Process->new();
my $peer = Net::BGP::Peer->new(
Start => 0,
ThisID => $myip,
ThisAS => $myas,
PeerID => $peerip,
PeerAS => $peeras,
KeepaliveCallback => \&sub_keepalive_callback,
UpdateCallback => \&sub_update_callback,
NotificationCallback => \&sub_notification_callback,
ErrorCallback => \&sub_error_callback,
OpenCallback => \&sub_open_callback,
ResetCallback => \&sub_reset_callback,
KeepAliveTime => 600,
HoldTime => 1800,
);
- Adjusting the BGP timers on the dynamips router, like so:
c7200# configure terminal c7200(config)# router bgp 65000 c7200(config-router)# timers bgp 600 1800 c7200(config-router)# end
We can then see those values reflected here:
c7200# show ip bgp neighbor 192.168.1.104 | in Last Last read 00:00:00, last write 00:00:56, hold time is 1800, keepalive interval is 600 seconds c7200#
Note that it’s probably best to change both sides (bgp_simple.pl and your router’s config), since BGP will use the lowest of the values configured between peers.
Tags: bgp, ccnp, cisco, internet, labs, linux, networking, open-source, osx, software | 6 Comments »
Configuring SNTP on ProCurve Switches
Written by jlgaddis on August 18, 2009 – 4:16 am -This is probably the shortest video I’ve ever made, but the task is really simple — configuring the Simple Network Time Protocol (SNTP) on an HP ProCurve switch.
For reference, here’s the “sntp” commands available in the software:
ProCurve Switch 2610-48(config)# sntp ?
broadcast Operate in broadcast mode
<30-720> The amount of time between updates of the system clock
via SNTP
server Configure SNTP servers to poll time from.
unicast Operate in unicast mode
<cr>
…and, of course, “timesync” takes an argument too…
ProCurve Switch 2610-48(config)# timesync ? sntp Set the time protocol to SNTP timep Set the time protocol to the TIME protocol
Pretty simple, huh? Here’s the video:
Tags: hp, labs, networking, security, switching, video | 1 Comment »
Video: Update ProCurve software via TFTP
Written by jlgaddis on August 17, 2009 – 9:02 am -Nothing earth-shattering here, but back in February, I posted “Upgrading ProCurve firmware via TFTP”. Since that time, I’ve started posting video demos of some of the things I write about, so I thought I’d make up a quick one for this task as well. I happen to have a 5412zl switch sitting here next to me that I just powered up for the first time, and I know from the search terms people use to reach my site that this is one of the most popular HP-related posts, so perhaps seeing this in action will help understand it as well.
The first thing you’ll want to do is head over to ProCurve’s “Software for switches” page and grab the appropriate software for your switch. At the time of this writing, K.13.63 is the latest available on the web site, however I grabbed K.13.68 from HP’s FTP server. Unzip the file that you download, read the release notes, and copy the software over to your TFTP server.
After that, it’s just a matter of following the steps in the previous post.
Here’s an update in action (note that this switch shipped with K.12.57 and all software versions after K.13.61 include a BootROM update, so we’ll our switch reboot automatically in the middle of the update. Also, by default, the switch use DHCP will acquire an IP address on VLAN 1, so we don’t have to configure one (though it’s always a good idea, of course).
Note that this whole process takes quite a bit longer than shown in the video. When I was done recording, the video came in at 08:19 long. By removing a lot of the “waiting”, I was able to get it down to the 02:42 that you see here.
Tags: hp, labs, networking, switching, video | No Comments »
Using BGP communities to influence routing, part 2
Written by jlgaddis on August 10, 2009 – 9:11 am -I wanted to follow-up to an earlier post, “Using BGP communities to influence routing”, and show some more ways we can use BGP communities to influence routing on the Internet. What we’re going to see today is very common in the real world, although it does take the cooperation of your service provider.
If you’re lucky, your service provider will have already published a listing of communities that they accept and what effect they have. The ones we’ll be using today come from an old XO Communications document that I found.
Here’s the topology we’re going to be working with today:
We’re going to be the customer, in AS 65001. To keep things simple, we’ll configure a single connection on one service provider, XO Communications in AS 2828. As you can see, XO peers with a number of other providers. We have three loopback interfaces that we’re going to create, and the IP addresses we assign to those will be the ones we’re advertising into BGP.
Under normal circumstances, XO would simply pass along our advertisements to its peers. We’re going to see how we change that, however, by applying some communities to our routes as we send them to XO.
Here we have a small example of the communities a service provider might accept. By tagging our routes with these values, we can control how XO handles those routes. For example, if we send community 2828:1003, XO won’t advertise that route to Sprint. Likewise, if we send community 2828:1308, XO will prepend its own AS number three times. The possibilities are endless.
The bottom half of that diagram shows what we’ll actually accomplish when we’re done. When advertising our routes, we’ll use communities to tell XO:
- don’t advertise 192.168.0.1/32 to Sprint (2828:1003),
- prepend twice when advertising 192.168.0.1/32 to Level3 (2828:1207),
- prepend once when advertising 192.168.0.1/32 to AT&T (2828:1108),
- prepend thrice when advertising 192.168.1.1/32 to Sprint (2828:1303),
- don’t advertise 192.168.1.1/32 to Level3 (2828:1007),
- prepend twice when advertising 192.168.1.1/32 to AT&T (2828:1208), and
- advertise 192.168.2.1 normally (no communities)
That’s a big list of demands, but it’s actually quite easy to do and — even better — it doesn’t require us to deal with another person at our service provider. By using these communities, we can do it ourselves (and change it whenever we want!).
For this lab, the XO, ATT, SPRINT, and LEVEL3 routers have already been configured. I’m intentionally not posting the configuration from XO, as it is out of the scope of this posting. While I may post it later, we’re only worried about how to configure things on the customer side.
Here we go…
Tags: bgp, cisco, labs, networking, video | 1 Comment »
Dynamips, a 7200, and a full BGP table
Written by jlgaddis on August 7, 2009 – 5:15 am -I’ve been using dynamips on and off for a while now and always knew it was cool as hell, but this just really impressed me:
c7200# show ip bgp summary BGP router identifier 192.168.1.201, local AS number 65101 BGP table version is 593019, main routing table version 593019 297882 network entries using 34852194 bytes of memory 297882 path entries using 15489864 bytes of memory 56505/56484 BGP path/bestpath attribute entries using 7006620 bytes of memory 51421 BGP AS-PATH entries using 1620448 bytes of memory 1632 BGP community entries using 107766 bytes of memory 0 BGP route-map cache entries using 0 bytes of memory 0 BGP filter-list cache entries using 0 bytes of memory BGP using 59076892 total bytes of memory BGP activity 298425/543 prefixes, 298425/543 paths, scan interval 60 secs Neighbor V AS MsgRcvd MsgSent TblVer InQ OutQ Up/Down State/PfxRcd 192.168.1.81 4 65001 592483 167 593019 0 0 01:19:55 297882
192.168.1.201 is a 7200 running under dynamips, and I just fed it a (nearly) full BGP table.
For the curious, that was the only router I had running under dynamips and the host is a MacBook (10.5.8), with a 2.2GHz Core2Duo and 4GB of RAM.
How fuckin’ cool is that!?
Tags: bgp, cisco, labs, open-source, osx | 8 Comments »
