A vSAN stretched cluster manages failures that occur due to the loss of a network connection between sites or the temporary loss of one site.

Stretched Cluster Failure Handling

In most cases, the stretched cluster continues to operate during a failure and automatically recovers after the failure is resolved.

Table 1. How Stretched Cluster Handles Failures
Type of Failure Behavior
Network Connection Lost Between Active Sites If the network connection fails between the two active sites, the witness host and the preferred site continue to service storage operations, and keep data available. When the network connection returns, the two active sites are resynchronized.
Secondary Site Fails or Loses Network Connection If the secondary site goes offline or becomes isolated from the preferred site and the witness host, the witness host and the preferred site continue to service storage operations, and keep data available. When the secondary site returns to the cluster, the two active sites are resynchronized.
Preferred Site Fails or Loses Network Connection If the preferred site goes offline or becomes isolated from the secondary site and the witness host, the secondary site continues storage operations if it remains connected to the witness host. When the preferred site returns to the cluster, the two active sites are resynchronized.
Witness Host Fails or Loses Network Connection If the witness host goes offline or becomes isolated from the preferred site or the secondary site, objects become noncompliant but data remains available. VMs that are currently running are not affected.