VMware Greenplum on VMware vSphere uses the three networks described below.
Internal network
The internal network is identified by the port group gp-virtual-internal
. It is used by Greenplum for internal communications. It requires a new VLAN ID and a number of non-routable contiguous static IP addresses that will be used for:
Determine the number of required IP addresses based on the number of coordinator hosts and segment hosts in your environment. For example, for a four ESXi host configuration, it requires 67 non-routable contiguous static IP addresses.
External network
The external network is identified by the port group gp-virtual-external
. It is used by Greenplum for external traffic through the coordinator and standby coordinator nodes. It requires a new VLAN ID and 4 routable static IP addresses that will be used for:
Load Backup and Restore network
The load backup and restore network is identified by the port group gp-virtual-etl-bar
. It is used by Greenplum for ETL traffic and for backup and restore traffic. It requires a new VLAN ID and a number of non-routable contiguous static IP addresses that will be used for:
Determine the number of required IP addresses based on the number of coordinator hosts and segment hosts in your environment. For example, for a four ESXi host configuration, it requires 67 non-routable contiguous static IP addresses.
Note that ETL or backup traffic must take place within the internal network gp-virtual-etl-bar
which is non-routable.
Check the Distributed Virtual Switch settings to make sure that the Maximum Transmission Unit (MTU) is set to 9000 Bytes:
The table below summarizes the distributed port groups that the Virtual Distributed Switch must have configured. Some of them might already be present in your VMware vSphere environment.
Port Group Name | Description |
---|---|
gp‑virtual‑internal | Used for Greenplum cluster internal communications, including interconnect, dispatch and mirroring. This network is usually air-gapped, and does not have internet connection. All IP addresses are statically assigned. |
gp‑virtual‑external | Used for Greenplum cluster user connections from outside the cluster. The Greenplum coordinator host exposes $PGPORT on this network. It is also used to connect to the coordinator host and standby coordinator host for DBA or troubleshooting purposes. This network usually has DHCP enabled. |
gp‑virtual‑etl‑bar | Used for ETL and backup/restore operations for the Greenplum cluster. |
VMware vSphere‑management | Used by vCenter to manage the ESXi hosts. It supports the VMkernel port of management. |
VMware vSphere‑vmotion | Used for VMware vSphere HA and DRS. It supports the VMkernel port of vMotion. |
vsphere‑vsan | Used for vSAN connections. It supports the VMkernel port of vSAN. |
vsphere‑vcenter | Dedicated management network, including vCenter, DHCP, DNS, and NTP services. Usually the IP addresses are statically assigned. |
vsphere‑vm | Used by non Greenplum virtual machines to connect to the company network. It usually has internet connectivity. This network usually has DHCP enabled. |
Create the above port groups in your VMware vSphere environment:
When creating and configuring the distributed port groups listed above, you must adhere to the following principles:
vsphere-vsan
must be always on its own active uplink.gp-virtual-internal
must be always on its own active uplink.gp-virtual-etl-bar
must be always on its own active uplink.Based on the principles above, and depending on your Dell EMC VxRail topology, the configuration of the port groups may vary. The next section describes the recommended topology of Dell EMC VxRail and documents how the port groups must be configured for this particular topology.
This Dell EMC VxRail topology consists of Dell EMC VxRail nodes with two 100GbE Network Cards each with two 100GbE ports. Two ports are connected to two TOR switches, and one optional connection to management switch for Integrated Dell Remote Access Controller (iDRAC). You can find more information about this topology on page 109 of the Dell EMC VxRail Network Planning Guide (Figure 53).
Configure the port groups in your VMware vSphere environment:
You must configure the following port groups on the Virtual Distributed Switch to match the following table:
Port Group Name | uplink1 (vmnic0) | uplink2 (vmnic1) | uplink3 (vmnic2) | uplink4 (vmnic3) |
---|---|---|---|---|
vsphere-management |
|
Unused |
|
Standby |
vsphere-vmotion |
|
Unused |
|
Active |
vsphere-vm |
|
Unused |
|
Active |
vsphere-vcenter |
|
Unused |
|
Active |
vsphere-vsan | Unused |
|
Standby |
|
gp-virtual-internal | Unused |
|
Active |
|
gp-virtual-external |
|
Unused |
|
Standby |
gp-virtual-etl-bar |
|
Unused |
|
Active |
For example, the Failover order configuration for port group gp-virtual-internal
should look like this:
Since the distributed switch and the physical switches are configured separately, settings such as the MTU, VLAN, and teaming settings might present configuration discrepancies. To help narrow down any network configuration issues, VMware vSphere provides a vDS Health Check to detect network configuration inconsistencies between the distributed switch and physical switches. Please refer to the VMware vSphere Documentation for more details.
To enable the VMware vSphere Distributed Switch Health Check from vCenter:
There are some known limitations of Health Check:
You may want to enable vDS Health Check only during deployment and testing phases, and turn it back off once you have verified that there are no configuration inconsistencies.
Continue configuring Dell EMC VxRail with VMware vSphere by Setting Up vSphere DRS and HA.