Create a Distributed Switch

Procedure

1. In the navigation pane, choose Inventory > Network Resource.
2. In the resource tree, right-click the target data center and select New Distributed Switch.
3. In the New Distributed Switch dialog, complete basic information configuration.

   Option	Description
   Name	Enter a name for the distributed switch.
   Description	Enter a brief description for the distributed switch.
   Data Center	Displays the data center where the distributed switch resides.
   Cluster	Select the cluster to attach to the distributed switch. Virtualization clusters and bare metal clusters are supported.

4. Complete network configuration as needed.

   Network configuration is required only when you select a virtualization cluster. Select an addition method based on your host NIC plan, and then configure the corresponding options.

   Add Individually or Batch Bonding

   Option	Description
   Addition Method	Select Add Individually or Batch Bonding.
   Uplink Name	Set a name for the bonded host physical NICs that connect to physical switches. By default, the name uses the Uplink+suffix format, and the suffix increments as 1/2/3. When the number of uplinks is 10 or more, the default name uses the Up+suffix format. A custom name must be 1 to 10 characters long, can contain only letters, numbers, hyphens, and underscores, and cannot start with a number.
   Bond Mode	Select a bond mode for physical NICs. In LACP mode (mode 4), bonded NICs share the same speed and duplex settings, and network traffic is distributed across the NICs. In active-backup mode (mode 1), the active NIC handles network traffic during normal operation and the backup NIC takes over automatically if the active NIC fails. Both modes support 1 to 8 physical NICs, and 2 or more NICs are recommended.
   Hash Policy	When LACP mode is selected, set a hash policy to determine the network traffic egress. layer 2+3: Picks out a NIC port to send data packets based on the hash computation on the source MAC address, destination MAC address, and IP address. layer 3+4: Picks out a NIC port to send data packets based on the hash computation on the IP address and port. TCP/IP stacks are supported. layer 2: Picks out a NIC port to send data packets based on the hash computation on the source MAC address and destination MAC address.
   Host NIC Ports	Select the physical ports on the host to bond. When creating bonds for each individual host, the selected ports on that host must operate at the same speed. When creating bonds for multiple hosts in bulk, you can only select ports that have the same speed.

   Specify the Same Port

   Option	Description
   Addition Method	Select Specify the Same Port.
   Port Type	Select the port type, including bond interfaces and non-bonded port.
   Bond/Host NIC Ports	Select a NIC to add all hosts in the cluster that have the same NIC.

5. Complete distributed port group configuration as needed.

   By default, a distributed port group is created. You can choose whether to create a distributed port group on the distributed switch.

   Option	Description
   Name	Enter a name for the distributed port group.
   VLAN Type	Select a VLAN type. If you select standard VLAN, enter the VLAN ID.
   DHCP Service	Select whether to enable the service that automatically assigns IP addresses to internal platform resources.
   DHCP Service IP	Set the IP address used by the DHCP service.
   DNS	Set the DNS resolution service for the distributed port group.

   IP Address Management

   After IP address management is enabled, you can add network ranges to the distributed port group. IP addresses in the network ranges can be assigned through DHCP to resources that use this network.

   IPv4 Address Type

   Option	Description
   IP Allocation Policy	After the DHCP service is enabled, the following allocation policies are supported: Random allocation: The system randomly assigns IP addresses in the added network range. Allocate in Order: The system assigns all available IP addresses in the added network range in ascending order. Released IP addresses are assigned in the next allocation. Example: Assume that the network range is 192.168.0.101～192.168.0.120, within which 192.168.0.101～192.168.0.108 are allocated. If 192.168.0.106 is released, it will be assigned first in the next allocation. Allocate in Cycle: The system assigns IP addresses in the added network range in ascending order. Addresses released during allocation are assigned after all currently available IP addresses have been assigned once. Example: Assume that the network range is 192.168.0.101～192.168.0.120, within which 192.168.0.101～192.168.0.108 are allocated. If 192.168.0.106 is released, it will be assigned after 192.168.0.120 is used.
   Network Range Method	The following network range methods are supported: IP Range: Enter the start IP address, end IP address, subnet mask, and gateway of the network range. Note: Do not include the gateway, broadcast address, or network address in the added IP range. CIDR: Enter the CIDR block and gateway. The gateway can be the first or last address in the CIDR block. If the gateway is left empty, the first address in the CIDR block is used by default.

   IPv6 Address Type

   Option	Description
   Network Range Method	The following network range methods are supported: IP Range: Enter the start IP address, end IP address, prefix length, and gateway of the network range. CIDR: Enter the CIDR block.
   IP Allocation Mode	Select the IPv6 address allocation mode: Stateful-DHCP: Stateful DHCP configuration. The interface address and other parameters are all configured through DHCP. The IP range method supports only this allocation mode. Stateless-DHCP: Stateless DHCP configuration. The interface address is derived from the prefix in router advertisements and the interface MAC address, and other parameters are configured through DHCP. SLAAC: Stateless address autoconfiguration. The interface address is derived from the prefix in router advertisements, and other parameters are also included in router advertisements.

6. After confirming that the configuration is correct, click OK.
   Note:

   • After a single NIC or bonded NIC is added to the distributed switch, bridge names created by distributed port groups use the br_dvs prefix.
   • If the management network address is on a NIC or sub-interface before the host is added, the management network address is moved to the br_dvs{ID}_{VLAN ID} bridge after the host is added. After the host is deleted, the management network address moves back to the NIC or sub-interface.
   • If the management network address is on a br bridge before the host is added, for example a user-defined br_{name}_{VLAN ID} bridge, the bridge name remains unchanged after the host is added. When the management network and service network share NICs, if the VLAN of the new distributed port group is different from the management network VLAN, the bridge created by the distributed port group still uses the br_dvs{ID}_{VLAN ID} format, but the NIC that carries the management network address is not affected.