DHCP command "client identifier x" ..... What do you really use

From: Smithson, Brandon K (brandon.k.smithson@citi.com)
Date: Mon Dec 31 2007 - 16:50:47 ARST


Is it really that simple to find the client identifier? Just issue the
debug ip dhcp server packet command and use the resulting displayed MAC
address?

I know the Cisco Univercd link
<<http://www.cisco.com/univercd/cc/td/doc/product/software/ios124/124cg/
hiad_c/ch10/hipdhcps.htm#wp1116280>> says to do that, but the DHCP
command "client identifier" is contradicted between that link and the
following Univercd link:
<<http://www.cisco.com/univercd/cc/td/doc/product/software/ios124/124cg/
hcf_c/ch15/cfhanstl.htm#wp1201022>>.

The second link, which pertains to Autoinstall, says you have to use the
MAC address and reference it to a matrix to determine the client
identifier (at least for the manual method). Regardless, my point is
that the end result is a long string, not just the MAC address. For
example the MAC address of R2 Ethernet 0 in the 2nd link above is
00e0.1eb8.eb09., and it's client identifier is
0063.6973.636f.2d30.3065.302e.3165.6238.2e65.6230.392d.4574.30

So, wouldn't you have to use (if the host is R2, and it's connected
interface is Eth0):
ip dhcp pool 1
   host 10.0.20.54 255.255.255.240
   client identifier
0063.6973.636f.2d30.3065.302e.3165.6238.2e65.6230.392d.4574.30

According to the second Cisco Univercd link above:
"Some Cisco networking devices use a DHCP client identifier format that
is different from the format used by networking devices running Cisco
IOS release 12.4(1) or newer. This document only explains the DHCP
client identifier format used by networking devices running Cisco IOS
release 12.4(1) or newer. Use the process described in "Determining the
Value for the DHCP Client Identifier Automatically" section to determine
the DHCP client identifier format that your Cisco networking device is
using."

To me, it reads as if you have to use the longer client id if your
devices are running 12.4(1) and newer. Since the lab devices are
running 12.4 and higher (I assume), do we use this more complex,
drawn-out method?

Thanks,
Brandon

-----Original Message-----
From: Darby Weaver [mailto:darbyweaver@yahoo.com]
Sent: Saturday, December 29, 2007 10:54 PM
To: Smithson, Brandon K [CCC-OT_IT]; Darren Johnson; Thorsten Mayr;
Wilson, Ryan # Atlanta; Cisco certification
Subject: RE: Lab Configuration Management

Yep:

This is the way I was taught to find it:

You can determine the client identifier by using the debug ip dhcp
server packet command. In the following example, the client is
identified by the value 0b07.1134.a029.

Router# debug ip dhcp server packet

DHCPD:DHCPDISCOVER received from client 0b07.1134.a029 through relay
10.1.0.253.

DHCPD:assigned IP address 10.1.0.3 to client 0b07.1134.a029.

However, in a typical DHCP Server (I'm an old Windows guy, so I'm going
to quote MS DHCP)...

I would just create static DHCP entries and for this I only need the mac
address of the interface I am expecting the DHCP request to come from.

But since we are Cisco guys and we know how to use debug effectively, we
can get the client identifier pretty quickly as well.

But since my TFTP Server is probably most of the times going to either
be a Windows or Linux box any way, I'd probably go ahead and configure
it to the be the DHCP Server in the first place.

--- "Smithson, Brandon K "
<brandon.k.smithson@citi.com> wrote:

> Thanks for the explanation. I believe "hardware address" is for BOOTP

> requests and "client identifier" is for DHCP requests.
>
http://www.cisco.com/univercd/cc/td/doc/product/software/ios124/124cg/hi
> ad_c/ch10/hipdhcps.htm#wp1074511
>
> The client identifier takes a little more work to calculate than just
> the hardware address.
>
http://www.cisco.com/univercd/cc/td/doc/product/software/ios124/124cg/hc
> f_c/ch15/cfhanstl.htm#wp1201022
>
>
> -----Original Message-----
> From: Darby Weaver [mailto:darbyweaver@yahoo.com]
> Sent: Saturday, December 29, 2007 5:32 PM
> To: Darren Johnson; 'Thorsten Mayr'; Smithson, Brandon K [CCC-OT_IT];
> 'Wilson, Ryan # Atlanta'; 'Cisco certification'
> Subject: Re: Lab Configuration Management
>
> Hmm...
>
> You guys ever look at the process of a Cisco IP Phone and how it grabs

> its configs?
>
> Well we used to do something amazingly similar with our WLSE and our
> WAPs just by dropping them on the correct VLAN...
>
> DHCP is a wonderful tool.
>
> Autoinstall is a wonderful tool.
>
> Doesn't hurt to be an expert with both of them for the Lab anyway.
>
> It really a concept, your router wants to load this way if it has no
> configuration...
>
> Minimal Cisco DHCP Server Configuration Example
>
> The following example shows the minimum configuration needed on a
> Cisco DHCP server to enable AutoInstall on a new router.
>
> ip dhcp pool 1
>
> host 10.0.20.54 255.255.255.240
>
> In this example the new router will download the default configuration

> file from any available TFTP server, using only broadcast TFTP
> requests.
> Configuring a Cisco DHCP Server to Provide TFTP Server Information To
> AutoInstall Examples
>
> In the following example, a router is configured as a DHCP server that

> will provide information to a new router. The new router is connected
> to the network via the Ethernet 0 interface, which has a MAC address
> of 0000.0c59.fcb0. The DHCP server is configured to assign the IP
> address of 10.0.20.54/28 to the Ethernet 0 (E0) interface on the new
> router.
> This configuration instructs the DHCP client (the new router using
> AutoInstall) to download the configuration file with the name
> "/tftpboot/R1-config" from the TFTP server with the address
> 172.16.1.1.
>
> ip dhcp pool 1
> host 10.0.20.54 255.255.255.240
> !The following line identifies the new router using the new router's
> MAC address
> hardware-address 0000.0c59.fcb0
> !The following line specifies the configuration filename
> bootfile R1-confg
> !The following line specifies the TFTP server address
> option 150 ip 172.16.1.1
>
>
> The following example configures the DHCP server to provide the same
> address to Ethernet 0, but in this example the TFTP server name
> "tftp.cisco.com" is given instead of the server address. The
> dns-server command is added to provide the address of the DNS server
> (173.1.1.10) that will resolve the TFTP server name.
>
> ip dhcp pool 1
> host 10.0.20.54 255.255.255.240
> hardware-address 0000.0c59.fcb0
> bootfile R1-confg
> !option 66 specifies the TFTP server name
> option 66 ascii tftp.cisco.com
> dns-server 173.1.1.10
>
> Configuring a Default Relay Router Example
>
> In this example the address 10.0.20.20 is specified as the next-hop
> toward the TFTP server with the address
> 172.16.1.1:
>
> ip dhcp pool 1
> host 10.0.20.54 255.255.255.240
> hardware-address 0000.0c59.fcb0
> bootfile R1-confg
> !option 150 specifies the TFTP server address
> option 150 ip 172.16.1.1
> default-router 10.0.20.20
>
>
> Option 66 and 150 is discussed here for instance.
>
> Note the BootFile Name and the R1Config (for
> instance)
>
> There is also an option to specify a unique file as well in DHCP and
> if you use this puppy you can have one file per router or switch for
> instance.
>
>
>
> This is not a totally new concept.
>
> Now once the router grabs a DHCP Address and Basic Config from
> Autoinstall....
>
> What else did you need it to do for you?
>
> When you finish the lab, you can have a preconfigured RANCID or Kiwi
> or AdventNet or Cirrus or CiscoWorks login and grab the configs for a
> quick compare aka grading.
>
>
>
> You can do something similar with a tool like SolarWind's Engineer's
> Toolkit and grab configs as well.
>
> But I like Rancid and Kiwi best. Of course RAT would do the same
> thing compare your completed config against a baseline configuration.
>
>
>
>
>
> --- Darren Johnson <dazza_johnson@yahoo.co.uk>
> wrote:
>
> > Hey all. Did the original post allude to the fact
> that this was for a
> > lab?
> > If so, to dynamically load a config onto a router,
> which is set to
> > factory default, you need to look at autoninstall.
> That would sort the
>
> > 'loading config' issue.
> > To dynamically save a config every specified
> interval (say Friday
> > morning at 8:00am to a TFTP server) you need to be
> looking at KRON.
> > Ive set this up for a customer recently (watch out
> for a software bug
> > when specifying a Sunday :-( .......)
> >
> > Dazzler
> >
> > -----Original Message-----
> > From: nobody@groupstudy.com
> > [mailto:nobody@groupstudy.com] On Behalf Of
> Thorsten Mayr
> > Sent: 28 December 2007 20:27
> > To: 'Smithson, Brandon K '; 'Wilson, Ryan #
> Atlanta'; 'Cisco
> > certification'
> > Subject: RE: Lab Config Management
> >
> > Just had a thought about this as I'll need a
> solution myself and a
> > click on a gui aka website sounds like a neat
> plan...
> >
> > The options I am seeing:
> >
> > a) go with something like "kiwi" where u can
> define multiple jobs to
> > archive the configs in different folders (as in
> various
> > labs...) and set up multiple
> > jobs to upload the configs the way round... or
> send static commands to
>
> > pull the config whatever you prefer.
> >
> > - problem you probably need a lic for kiwi or is
> there a freeware
> > version - and kind of generic solution...
> >
> > b) write a simple script to push and pull the
> configs which u can bind
>
> > to a web front end easily. Easy to do, but
> initially intense to write
> > all those scripts
> >
> > - don't like the generic side of this idea
> neither...
> >
> > c) write a php script with arrays, something like
> $devs =
> > array('router1' => array( 'hostname' =>
> '192.168.0.1' ....
> > And /path/to/somewhere/$LAB_ID/$hostname etc...
> >
> > - somehow I, well let's say "don't like" php ...
> ):
> >
> >
> >
> > My 2 cents - if someone has a solution in place
> already - please
> > please share :)
> >
> > ... I checked on freshmeat but couldn't find
> anything (yet)
> >
> >
> > -----Original Message-----
> > From: nobody@groupstudy.com
> > [mailto:nobody@groupstudy.com] On Behalf Of
> Smithson, Brandon K
> > Sent: 28 December 2007 18:30
> > To: Wilson, Ryan # Atlanta; Cisco certification
> > Subject: RE: Lab Config Management
> >
> > You could do a term len 0, show run, highlight the
> config, copy and
> > paste it to Wordpad (or Word if you have it). If
> you have large
> > configs, you can set a send delay (around 10 ms)
> to slow down copying
> > the large configs back into the routers to prevent
> overflowing and
> > causing errors.
> >
> > -----Original Message-----
> > From: nobody@groupstudy.com
> > [mailto:nobody@groupstudy.com] On Behalf Of
> Wilson, Ryan # Atlanta
> > Sent: Friday, December 28, 2007 10:46 AM
> > To: Cisco certification
> > Subject: Lab Config Management
> >
> > I was wondering if someone could lead me in the
> right direction. I
> > have a LAB and I would like the ability to save
> and load my config
> > files dynamically. I have seen this done in class
> room environments,
> > but I'm not how. Does anyone know of a solution?
> >
> >
> >
> > Ryan
> >
> >
>



This archive was generated by hypermail 2.1.4 : Tue Jan 01 2008 - 12:04:32 ARST