vSAN -- Health Service -- Proactive Tests -- Network Performance Test
Article ID: 326905
Updated On:
Products
Issue/Introduction
This article explains the Proactive Test "Network performance test" in vSAN (formerly known as "Multicast performance test" in releases before vSphere 6.5 P02) and provides details on why it might report an error.
The Network performance test can be used to assess if there are any connectivity issues in the network used by vSAN. In addition, the result can be used as a reference to check if the network speed among the hosts meets the hardware specification and is sufficient to support the workloads hosted on vSAN.
In a regular or stretched cluster configuration, the test is performed between the hosts in the cluster. However, if HCIMesh is in use and the test is triggered on the client cluster, it will verify the network bandwidth between the hosts in the client cluster and the hosts in the server cluster separately.
During the test, all connected hosts are organized to form a circular list, and every host will be sending network packets to the next host while receiving packets from the previous host at the same time. Upon the completion of the test, the actual network bandwidth of each link between the sender host and the receiver host is reported.
testID: com.vmware.vsan.health.test.networkperftest
To access this proactive network performance test, go to: Web Client --> Select your vSAN Cluster --> Monitor Tab --> vSAN --> Proactive Tests
This test is part of vSAN’s proactive tests, for more information, please check Proactive Test document.
Environment
VMware vSAN 7.0.x
VMware vSAN 6.7.x
Resolution
Q: How to determine if there are network connectivity issues?
A: There are 2 things to be considered when interpreting the test result:
1) does the result meet the hardware specification?
2) is the network speed sufficient to support the workloads hosted on vSAN?
vSAN requires a minimum NIC speed of 1Gbps per host, and vSAN ESA requires a minimum aggregated NIC speed of 25Gbps per host. The bandwidth required for vSAN is decided by the workloads hosted on vSAN, and in general more bandwidth is required to support more vSAN workloads. Both the result state and the result value shall be used to determine if the network connection is at a healthy state. Take an example, a vSAN cluster configured 10Gbps for vSAN on each host, the result state shows all green and the result value is somewhere around 5Gbps, even though the test state doesn’t report any error, but there could still be network issues in the cluster as the test cannot achieve a result close to 10Gbps.
The test will show a resulting state:
For vSAN:
- Red if the speed is 0 Mb/s
- Yellow if the speed is less than 850 Mb/s
- Green if the speed is greater than or equal to 850 Mb/s
For vSAN ESA
- Red if the speed is 0 Mb/s
- Blue if the speed is greater than 0 Mb/s
Note: Any network connection that fails to meet the hardware specification should be checked to ensure the network bandwidth will not affect vSAN performance.
Q: How to troubleshoot and fix the network connectivity issues?
A: Below is a list of the most commonly seen errors that could cause network connectivity issues.
- Uncertified or outdated NIC driver/firmware
- Incorrect network connection/cabling for NIC, switch port, etc
- Incorrect virtual switch/port group/VLAN setting
- Incorrect MTU setting on the physical switch, DVS or vmk
- Improper TCP ring size setting
- Malfunctioning hardware or not certified switch vendor cable choice
- Insufficient physical switch/port buffer
- Over-committed or insufficient designed ISL between switch/switch layers
Note: in case you have difficulty identifying the root cause of the problem please reach out to your field support for troubleshooting assistance.
Q: How to design the network to achieve the best performance of vSAN?
A: Please go to vSAN Network Design guide for guidance.
Feedback
