PowerFlex: SVMs running on ESXi using Distributed Virtual Switches (dvSwitch) will not reconnect automatically after DU network outage
Summary: When recovering from a DU caused by a network outage, SVMs might not reconnect back automatically even if network connectivity has been restored - system remains in DU state.
Symptoms
When recovering from a DU caused by a network outage ESXi hosts running SVMs remain disconnected from vCenter. And may not reconnect back automatically even if network connectivity has been restored - system remains in DU state.
HCI vSphere environment dvSwitches are in use. A network outage causes a DU, ESXi hosts disconnect from vCenter. After the network problem is fixed, SVMs will not reconnect automatically, DU continues.
Cause
When using dvSwitches, ESXi hosts must remain connected to vCenter. During DU, due to the way ESXi handles the APD situation, hypervisors disconnect from vCenter which makes vdSwitches unusable. The only way to restore connectivity to vCenter (and in effect to restore vdSwitches) is to reboot each ESXi host which disconnected from vCenter.
Resolution
This is a known issue and has been described in many VMware KB articles.
ESXi hosts in All Paths Down (APD) condition may appear as Not Responding in vCenter Server
Permanent Device Loss (PDL) and All-Paths-Down (APD) on host
It is not related to VxFlex OS, but it affects its operations and recovery capabilities.
Workaround: Reboot each ESXi host which disconnected from vCenter. You might also consider enabling PDL feature on the ESXi side to prevent them from disconnecting from vCenter.
Additional Information
Impacted Versions
All versions - this is ESXi behavior, not a VxFlex OS problem.