ESXi Hosts show alert: “quick stats not up to date” when consolidating snapshots on ESXi 7.0U2 and above while using SIOC policies with limits and shares below 400.

Products

• VMware vSphere ESXi

Issue/Introduction

Symptoms:

• Snapshot consolidation fails and causes the host to go unresponsive
• hostd logs may show entries similar to:

[YYYY-MM-DDTHH:MM:SS] warning hostd[2515463] [Originator@6876 sub=IoTracker] In thread 2515941, preadv("/vmfs/devices/vdfm/1088ff3-vdfm") took over 840 sec.

• The ESXi host is on ESXI 7.0U2 or above
• An SIOC policy is used on the VM being consolidated with shares at or below 400

Environment

VMware vSphere ESXi 7.0.2

Resolution

This issue is resolved in vSphere ESXi 7.0 U3l (build 21424296).
This issue is resolved in vSphere ESXi 8.0b (build 21203435).

Workaround:
To resolve the issue with the unresponsive host, reboot the host.

Disabling SIOC policies or increasing the Shares and Limits will allow consolidation to work. 
In addition, cloning the VM in question resolves the issue. 

Furthermore, the issue is not seen when SPM is disabled. The issue is also not seen with policies with higher limits such as Gold or when the lead policy is reapplied.

Additional Information

Impact/Risks:
After upgrade to ESXi 7.0U2 or above, environments that utilise limits/shares below 400 on SIOC policies will begin to see consolidation tasks that cause the host to go into a hung state.