I am a little confused by the entries below. Keep in mind that your catalyst 1900 and 3500 switches are DHCP clients right out of the box. All you have to do is plug them in and get connectivity (somehow) to a DHCP server and it will go through DHCP discovery and lease an address. Somehow however, that's not exactly what I believe you were really trying to reference. If what you really meant to say was that clients connected on individual ports are not able to obtain a lease, that is a different matter. In order for a DHCP client to operate effectively in a switched internetwork, you need to have the following tasks completed: 1. A properly configured client. This means that all NIC drivers are properly installed and all Windows settings for Gateway, IPaddr, Mask, WINS, etc. have been cleared and removed from the network settings. 2. A properly configured switch and switch port. I count these as separate animals. For example, a properly configured switch has trunking set up to the distribution switch and Fast EtherChannel appropriately configured on the trunk links(if applicable). All devices that use that switch will depend upon the trunk link being properly configured and optimized. Speed and duplex should definitely be hard coded to ensure proper operation (FD and 100MBPS/1000MBPS). Any VLANs that should be excluded from the trunk, should be removed. Finally, all administrative data should be set for the switch, such as hostname, IPaddr/mask(hard code this; do not use DHCP), Default Gateway on the management VLAN, basic security measures and uplinkfast (if appropriate). 3. A properly configured access port on the above switch will also have speed and duplex hard coded (if appropriate). It is appropriate if the host does not move, or has a name of "server". If it is a port used by road warriors, then auto speed and duplex may be a necessity and YMMV. An absolute necessity is to turn off trunking, turn off bundling/port channeling (FEC and GEC), and set spanning-tree portfast to on. For CAT5ks running 6.1 code and CAT6ks, this can be done using the "set port host" command. For your lower end IOS based switches, you will need to apply/remove each of the individual commands. This should ensure smooth operation of access ports and the previous paragraph should ensure smooth connectivity back to the distribution switch. All of these configuration tasks are dealing with *layer 2* connectivity issues. 4. Assuming your hosts are on separate VLANs, ensure that the access port is assigned to the proper VLAN. Just remember, VLANs are a layer 2 thing. Subnets are a layer 3 thing. Make the mental glue for both; VLAN=Subnet=Broadcast domain. Don't put the right host in the wrong VLAN. 5. At your distribution layer, you will need to configure routing and your router. Now you are reaching up to layer 3. You will need to configure IP addresses on each VLAN interface on an RSM, or subinterfaces on an external router. Both can be used, but an internal RSM tends to be more convenient (and compact). When you configure each of the VLAN interfaces, you will need to specify an IP helper address command. This should be done on each VLAN interface. If you are using multiple DHCP servers, you may wish to split up your scopes to achieve some form of load balancing (unless you are suing CNR) Additionally, you can use a helper address that forwards to a network address or a host address. I would exercise caution in using a network address, because I remember a consulting issue where this resulted in a routing loop. In addition, do not forget the use of the "ip forward-protocol" command to limit which of the seven protocols get forwarded to the helper address. In a MS environment, this is an absolute necessity. 6. Ensure you have a path to your DHCP server from all points in the network. This implies that routing is up and running and properly configured and fully converged. Remember, the DHCP server will give you the "correct" IPaddr, GW, mask, and other info based upon the GIaddr. The GIaddr is that Gateway Interface address that exists on each of those VLAN IP addresses on your router/RSM. Avoid the use of secondary addresses on VLAN interfaces. 7. Configure your DHCP server properly to add scopes for each of the VLANs you have created. I will leave this exercise up to you on how to properly configure the DHCP server of your choice. 8. Implement and troubleshoot your config. What works and why? What doesn't work and why; what is different from the devices that do work? Always look to spot configuration errors early on. Use utilities such as ping and trace to verify layer 3/4 connectivity. If all else fails, check cabling. To quote a good friend, "the root of all evil in networking is cabling." Good luck! HTH, Paul Werner > Hi all, > I have a question concerning Catalyst switches. I have two 3500's and > one > 1900 and when I try to release and renew an IP address on the 1900 it is > fine, but when I try on the 3500 it appears to drop the request. Is > there > an IP-HELPER address on the switch such as there is on a router? ________________________________________________ Get your own "800" number Voicemail, fax, email, and a lot more http://www.ureach.com/reg/tag _________________________________ FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
