Slow Network Data Transfer Between Virtual Machines in Different Datacenters
search cancel

Slow Network Data Transfer Between Virtual Machines in Different Datacenters

book

Article ID: 444322

calendar_today

Updated On:

Products

VMware vSphere ESXi

Issue/Introduction

Virtual machines experience slow network data transfer or high latency when communicating across different datacenters.

Cause

A physical network constraint or latency bottleneck on the external infrastructure or circuit connecting the datacenters, rather than a performance degradation within the ESXi host or virtual machine.

Resolution

  • Verify Traffic Shaping is disabled on the relevant vSphere Distributed Switch (vDS) or Standard Switch portgroups via the vCenter Server UI.

  • Establish an SSH session to the ESXi host where the affected virtual machine resides.

    • Run esxtop, press n to access the networking view, and verify that the %DRPTX and %DRPRX columns display 0.00 to confirm the virtual machine is not actively dropping packets.

    • Identify the active uplinks for the virtual switch and run esxcli network nic stats get -n vmnicX (replacing X with the vmnic number). Verify there are no actively incrementing dropped packets or receive missed errors.

  • If ESXi network statistics show no drops, perform bandwidth baseline testing using iperf between two virtual machines residing on the same ESXi host.

    • iperf is a tool for active measurement of the maximum achievable bandwidth on IP networks.

    • Perform iperf testing between virtual machines on different ESXi hosts within the same datacenter to validate local physical switch performance.

    • Perform iperf testing between the virtual machines across the different datacenters.

    If the bandwidth drops significantly only during the inter-datacenter test, engage the physical network team to inspect the interconnecting circuit.

Additional Information