HP Procurve uses the same command, IP helper-address 111.222.333.444 as Ron mentioned is used for Cisco IoS. Cisco SMB switches like the SG300 use ip dhcp relay address 111.222.333.444. The list is endless really. That's not my purpose. I just wanted to add the "why" you need to use this command if you want to use one DHCP server for multiple VLANs/subnets.
Single DHCP Server with Multiple VLAN pools
We're doing this with Server 2003 where I am. The key is our central layer 3 device (in our case an old 3Com 4900 SX). It has what 3Com calls a "virtual interface" for each vlan. When dhcp requests arrive at the device from the outside it forwards them using same vlan interface they arrived on. The dhcp server just has a scope setup for each vlan with no other special configuration details. It matches the dhcp request to the originating interface on the layer 3 device to know which scope to use. That's out of the box behavior for the windows dhcp server.
The downside to this approach is that segment with your server must now be a member of all vlans (on the same virtual segment). This means all your broadcast traffic will now go up the fiber link to your main office (though not down all the other links to other branches), and it's nice to keep those links as clean as possible. It'll also use time on what is likely already your busiest local segment to reach your server. But this is probably unavoidable anyway.
While there may be reasons that you use multiple physical uplinks with untagged frames for communicating with different VLANs, a more efficient approach is to use Tagged VLAN frames on a single network connection. When you use tagged frames, the Windows Server tags the frames appropriately for each VLAN it is associated with. The ability to Tag frames generally needs vendor-specific driver sets to be loaded, such as the Intel Pro Set drivers using Intel-based cards.
The principles still apply to Windows Server DHCP servers running inside a Windows Server virtual machine. The one method that is more difficult is the tagging of a single interface with multiple VLANs as this is not a feature that you can carry out with VMware Tools drivers that I am aware of. You can add multiple virtual NICs to a single VM and connect the virtual machine to different VLAN-backed port groups for each connection.
DHCP server and DHCP client sould be in the same vlan to be able to communicate as the initial DHCP discover is a layer 2 broadcast packet to ff:ff:ff:ff:ff:ff MAC address. Cisco routers and layer 3 switches are able to act as DHCP relay and forward DHCP requests to a DHCP server located in another VLAN : a single DHCP server can now be deployed to deliver IP addresses to many subnet.
I have configured 3 VLANs: 1 (default), 2 and 3. They're dedicated to 3 separate /24 networks. Host on these VLANs are able to communicate with each other, and each subnets having their own DHCP servers, working as expected. I want to simplify and unify the DHCP administration to just use a single DHCP server, specifically the DHCP server of the switch. I have the DHCP server enabled with 3 address pools. I also created VLAN interfaces in each VLAN/subnet (not sure if this is required for DHCP to work in different subnets).
othohin - Your response would be true if you only wanted one VLAN to appear to your ESXi servers. Instead what you need to do is configure the network ports which are being used with the ESXi server running View to support "trunks" and not "access" VLANs. If you wanted to have say VLAN 10,11,12 visable to the ESXi server you would setup the trunk on all the ports used by the servers and allow all 3 vlans. Now in ESXi under "Network" configuration you would add a "Vitrual Machine" for each of the 3 VLAN, one for each of the vlan you setup above. Now when you setup our desktop image for View make sure on it's network configuation that you select the proper Vitural Machine and it and any replicans from that image will now be attached to the selected VLAN. If you want images running on more VLAN than you must have an image for each VLAN.
Same thing happends in View..... When you setup your Desktop pool you will select a Default Image to use from your ESXi servers and when View generates desktops in the pool the network setting you had in the Default Image will be cloned to very desktop in the pool and which Vitrual Switch is being used and hence which VLAN is "TAGGED". If you need a different vlan tagged on the desktops in a pool you must then use a different default image with the proper network settings.
In this LAB, I am going to share with us on how to configure DHCP servers for VLANs in router on a stick scenario. Three sub-interfaces will be created on the router, each representing a VLAN, with each sub-interface having a dhcp server configured to handle IP address leasing to hosts in that VLAN. Though I have treated VLAN setup before, find it here, to ensure that everyone is carried along, this LAB will include vlan creation.
The objective of this Lab is to configure multiple dhcp servers for vlans on the router to service hosts in different vlans. We will start by configuring sub-interfaces with 802.1q encapsulation on the router and the assignment of IP addresses from different subnets to the sub-interfaces. Then we go over to the switch and create vlans. Finally, we will wrap things up with the configuration of dhcp servers for vlans on the router. Hosts in a vlan will acquire IP addresses from a dhcp server configured for that vlan.
I managed without problems to configure the static DHCP pools for the first 16 server, but I am unable to do so for a further one (I have 20 servers each rack). When I try to create a new DHCP pool I get:
Step 3.We enable DHCP relay and define pools and DHCP relay source and destination. Where source is switches VLAN IP and destination DHCP server IP. Finally we save configuration.Switch(config)#service dhcpSwitch(config)#ip dhcp relay information optionSwitch(config)#ip dhcp pool corporateSwitch(config-dhcp-pool)#relay source 192.168.10.90 255.255.255.0Switch(config-dhcp-pool)#relay destination 192.168.1.100Switch(config-dhcp-pool)#exitSwitch(config)#ip dhcp pool guestSwitch(config-dhcp-pool)#relay source 192.168.20.90 255.255.255.0Switch(config-dhcp-pool)#relay destination 192.168.1.100Switch(config-dhcp-pool)#Switch#copy run start
In summary, on LAN >> General Setup >> Details page, we can configure the TCP/IP settings as well as the DHCP server options. Vigor Router is a DHCP server by default, and it can be turned off or be a DCHP relay agent to forward the DHCP request to another DHCP server. Vigor Router is compatible with 802.1Q VLAN, we can have either port-based VLAN or tag-based VLAN setup on the router, and the router provides multiple subnet which allows us to have different TCP/IP settings and DHCP server configuration for each VLAN.
DHCP in a Windows failover cluster. This option places the DHCP server in a cluster with an additional server configured with the DHCP service that assumes the load if the primary DHCP server fails. The clustering deployment option uses a single shared storage. This makes the storage a single point of failure, and requires additional investment in redundancy for storage. In addition, clustering involves relatively complex setup and maintenance.
For BIG-IP systems with ePVA hardware support, the system includessupport for the IEEE 802.1QinQ standard. Known informally as Q-in-Q ordouble tagging, this standard provides a way for you to insert multipleVLAN tags into a single frame. This allows you to encapsulatesingle-tagged traffic from disparate customers with only one tag.
A common use case is one in which an internet service provider creates asingle VLAN within which multiple customers can retain their own VLANswithout regard for overlapping VLAN IDs. Moreover, you can usedouble-tagged VLANs within route domains or vCMP guests. In the lattercase, vCMP host administrators can create double-tagged VLANs and assignthe VLANs to guests, just as they do with single-tagged VLANs. For avCMP guest running an older version of the BIG-IP software,double-tagged VLANs are not available for assignment to the guest.
Subnetting allows you to create multiple logical networks that existwithin a single Class A, B, or C network. If you do not subnet, you areonly able to use one network from your Class A, B, or C network, whichis unrealistic.
Routing is the process of selecting a path for traffic between networksor across multiple networks. If a network device needs to communicatewith another device off of its local area network (LAN), it must know away to get there. For a node with a single network interface on a LAN,it will normally just be to the router acting as a default gateway forthe IP subnet. But some network devices have multiple network interfaceson different networks to use or have the responsibility of routingtraffic in the network and the next hop to get to a particulardestination may be out a particular interface. The list of which way togo to get to the different destination IP addresses is known as a routetable. Static routes can be added to a systems route table which willallow it to know which way to go to get to all of the IP subnets in anetwork, but the management of such a large table would be daunting andvery inefficient. And if a device failed in the path to a destination,the system would lose communications with the destination until the pathrecovered or was manually edited by an administrator to use a differentpath that may be available. This is where dynamic routing attempts tosolve the administrative inefficiency problem by constructing routingtables automatically, based on information learned by routing protocols.This allows devices on the network to act nearly autonomously inavoiding network failures and blockages. 2ff7e9595c
Comentarios