Microsoft Windows Server Failover Clustering (WSFC) with shared disks on VMware vSphere 7.x: Guidelines for supported configurations
search cancel

Microsoft Windows Server Failover Clustering (WSFC) with shared disks on VMware vSphere 7.x: Guidelines for supported configurations

book

Article ID: 313230

calendar_today

Updated On:

Products

VMware vSphere ESXi

Issue/Introduction

VMware vSphere 7.x provides flexibility and choice in architecting high-availability solutions using Windows Server as guest Operating Systems (OS). VMware vSphere 7.x supports Windows Server Failover Cluster (WSFC) with shared (clustered) disk resources by transparently passing to the underlying storage or emulating on the datastore level SCSI-3 Persistent Reservations (SCSI3-PRs) SCSI commands, required for a WSFC node (VM participating in a WSFC, further references as VM node) to arbitrate access to a shared disk. It is a general best practice to ensure that each node participating in a WSFC has the same configuration. 

Note: Clustered (“shared”) VMDKs are now available with vSphere 7.0

Information in this article is applicable to configurations when VMs hosting nodes of a WSFC (VM nodes) are located on different ESXi hosts – known as “Cluster-across-box (CAB)”. CAB provides High Availability (HA) both from the in-guest and vSphere environment perspective. VMware does not recommend configurations, where all VM nodes are placed on a single ESXi host (so called “cluster-in-a-box”, or CIB). The CIB solution should not be used for any production implementations – if a single ESXi host will fail, all cluster nodes will be powered off and, as the result, your application will experience downtime.
This KB assumes that all components underneath the WSFC-based VMs are architected to provide the proper availability and conform to Microsoft’s supportability as it relates to the in-guest configuration of this article. 

Note: A single WSFC consisting of both physical nodes and virtual machine is supported. For more information, see Cluster Physical and Virtual Machines section in the Setup for Windows Server Failover Clustering.

Environment

VMware vSphere 7.0.x

Resolution

This article provides guidelines and vSphere support status for guest deployments using Microsoft Windows Server Failover Clusters (WSFCs) with shared disk resources across nodes in CAB configuration on VMware vSphere 7.x.

vSphere Version Support for WSFC

Table 1 shows supported versions of Windows OS and VMware vSphere, being qualified by VMware. VMware does not impose any limitations nor requires a certification for application using WSFC on a supported Windows platform, therefore, any application running on a supported combination of vSphere and Windows OS is supported with no additional considerations. 

Note: Other WSFC-based solutions, not accessing shared disks (Microsoft SQL Server Always on Availability Groups (AGs) or Microsoft Exchange Database Availability Groups (DAG) require no special storage configurations on the vSphere side (VMFS or NFS). This kb should not be used for such configurations.

Table 1. Versions of Windows Server Supported by vSphere for a WSFC
 

Windows Server Version1

Minimum vSphere version

Maximum Number of WSFC Nodes with Shared Storage Supported by ESXi
2022 vSphere 7.0 5

2019

vSphere 7.0

5

2016

vSphere 7.0

5

2012 / 2012 R2

vSphere 7.0

5

  1. SQL Server 2016, 2017 and 2019 Failover Cluster Instances (FCI) were used to validate the WSFC functionality on vSphere and Windows Server versions listed in this table. 

  2. If the cluster validation wizard completes with the warning:” Validate Storage Spaces Persistent Reservation,” you can safely ignore the warning. This check applies to the Microsoft Storage Spaces feature, which does not apply to VMware vSphere.
     

VMware vSphere features support for WSFC Configurations with shared disks

The following VMware vSphere features are supported for WSFC:


The following VMware vSphere features are NOT supported for WSFC:

  • Live Storage vMotion.
  • Fault Tolerance (FT).
  • N-Port ID Virtualization (NPIV).
  • Mixed versions of ESXi hosts in a vSphere cluster in production use.
     

Support requirements for vMotion of a VM hosting a node of WSFC

Live vMotion (both user- or DRS-initiated) of VM nodes is supported in vSphere 7x under the following requirements: 

  • The VM virtual hardware version (“VM compatibility”) must be version 11 (vSphere 6.0) or later.
  • The value of the SameSubnetThreshold Parameter of Windows cluster health monitoring must be modified to allow 10 missed heartbeats at minimum. This is the default in Windows Server 2016. This recommendation applies to all applications using WSFC, including shared and non-shared disks.
  • Must configure cluster nodes with a DRS Affinity rule to prevent hosting more than one node of a WSFC cluster on a single ESXi host. This means that you must have N+1 ESXi hosts, where N is the number of WSFC nodes.
  • The vMotion network must be using a physical network wire with a transmission speed 10GE (Ten Gigabit Ethernet) or more. vMotion over a 1GE (One Gigabit Ethernet) is not supported.
  • Shared disks resources must be accessible by the destination ESXi host.
  • Supported version of Windows Server used (see Table 1 for the list of supported versions of Windows Server).

Storage Configuration

vSphere options for presenting a shared storage to a VM node are shown in the Table 2. 
Table 2. Supported storage configuration options

vSphere version

Shared disk options

SCSI bus sharing

vSCSI Controller type

Storage Protocol / Technology

vSphere 7.0

Clustered VMDKs

physical

VMware Paravirtual (PVSCSI), LSI Logic SAS

FC

vSphere 7.0

VMware vSphere® Virtual Volumes (vVols), RDM physical mode

physical

VMware Paravirtual (PVSCSI), LSI Logic SAS

FC, FCoE, iSCSI

VMware Cloud on AWS

Clustered VMDKs

physical

VMware Paravirtual (PVSCSI), LSI Logic SAS

N/A / vSAN

vSAN (vSphere 7.0)

Clustered VMDKs

physical

VMware Paravirtual (PVSCSI), LSI Logic SAS

N/A / vSAN


Clustered VMDKs

Requirements
VMware ESXi, VMware vCenter®, VMware vSphere VMFS

  • Require all hosts connected to a clustered VMDK datastore to be on ESXi version 7.0 or higher, managed by the same vCenter instance, version 7.0 and higher while disabling or enabling clustered VMDK flag on the datastore. Once the clustered VMDK flag is enabled/disabled, hosts can be managed by any vCenter with version 7.0 or later.
  • All ESXi hosts involved in hosting nodes of a WSFC must be managed by the same vCenter instance. A cross vCenter WSFC is not supported (i.e., when ESXi hosts, hosting VM nodes of a WSFC are managed by different vCenter instances)
  • Requires VMFS version 6. 

Virtual Machine (VM)

  • VMDKs must be Eager Zeroed Thick (EZT) Provisioned.
  • Clustered VMDKs must be attached to a virtual SCSI controller with bus sharing set to physical. VM Boot disk (and all VM non-shared disks) should be attached to a separate virtual SCSI controller with bus sharing set to none. Mixing clustered and non-shared disks on a single virtual SCSI controller is not supported.
  • Multi-writer flag must NOT be used.

 

Datastore

  • Require the datastore capability Clustered VMDK support to be enabled.
    Note: Perennial reservations are required on all datastores with Clustered VMDK support enabled.

Storage Array/LUN

  • Support SCSI-3 reservation type WEAR (i.e., write exclusive all registrant).
  • Supported storage system access protocol – Fiber Channel (FC) only.
  • The physical disk on which VMDKs are stored should support ATS commands.
  • Require 512/512e sector size disks.
  • Storage devices can be claimed by VMware Native Multipathing Plugin (NMP) or any other third party (non-vmware) Multipathing Plugins (MPPs). But please check with vendor regarding the support for Clustered VMDK before using their plugin (MPP)      

         See Clustered vmdk datastore LUN check.

 

Supportability

  • Maximum number of 128 clustered VMDKs per ESXi host.

  • Supports up to three (3) WSFC clusters (i.e., multi-cluster) running on the same ESXi host.

  • Mixing of clustered VMDKs and other types of clustered disks (e.g., pRDMs, vVol) in a single VM is not supported.

  • Placing all VMs, nodes of a WSFC on the same ESXi host (i.e. Cluster-in-a-Box (CiB) is not supported. 

  • VMs, nodes of a WSFC, must be placed on different ESXi hosts (i.e. Cluster Across Boxes (CAB)). The placement must be enforced with DRS MUST Anti-Affinity Rules. 

  • Change/increase in the GOS the WSFC Parameters.

    • (get-cluster -name <cluster-name>).QuorumArbitrationTimeMax = 60

    • (get-cluster -name <cluster-name>).SameSubnetThreshold = 10

    • (get-cluster -name <cluster-name>).CrossSubnetThreshold = 20

    • (get-cluster -name <cluster-name>).RouteHistoryLength = 40

  • Only datastores accessible via FC are currently supported. 

  • Non-SCSI back-end or non-SCSI virtual front-end (e.g., NVMe, vNVMe) are not supported.

  • SCSI-2 Reservations will not be supported on clustered disks/datastores.

RDM Configuration

RDMs used as clustered disk resources must be added using physical compatibility mode.
Virtual SCSI Controllers

  1. Mixing non-shared and shared disks. 

Mixing non-shared and shared disks on a single virtual SCSI adapter is not supported. For example, if the system disk (drive C:) is attached to SCSI0:0, the first clustered disk would be attached to SCSI1:0. A VM node of a WSFC has the same virtual SCSI controller maximum as an ordinary VM - up to four (4) virtual SCSI Controllers. 

  1. Modify advanced settings for a virtual SCSI controller hosting the boot device.

Add the following advanced settings to the VMs node: 

  • scsiX.returnNoConnectDuringAPD = "TRUE" 

  • scsiX.returnBusyOnNoConnectStatus = "FALSE" 

Where X is the boot device SCSI bus controller ID number. By default, X is set to 0.

  1. Virtual disks SCSI IDs should be consistent between all VMs hosting nodes of the same WSFC. 

  2. vNVMe controller is not supported for clustered and non-clustered disk (for example, a boot disk must NOT be placed on vNVMe controller). Check this Configuring disks to use VMware Paravirtual SCSI (PVSCSI) controllers for more details on how to change a controller for the boot disk: 

  3. Multi-writer flag must NOT be used.

  4. Use VMware Paravirtual virtual SCSI controller for the best performance.

For the best performance consider distributing disks evenly between as much SCSI controllers as possible and use VMware Paravirtual (PVSCSI) controller (provides better performance with lower CPU usage and is a preferred way to attach clustered disk resources).

Storage protocols

  • FC, FCoE and Native iSCSI are fully supported with pRDMs and vVols.
  • Clustered VMDKs are supported with FC only.
  • NFS is not a supported storage protocol to access a clustered disk resource for WSFC. NFS backend VMDKs can be used as non-shared disks (system disk, backup, etc.) without limitations.
  • vSAN supports natively clustered VMDK in vSphere 7.x
  • Virtual Volumes – supported. Check with your storage vendor if the implementation of vVol includes support for WSFC on VMware vSphere.

Multipathing configuration: Path Selection Policy (PSP)

  • RDM physical mode, vVols, Clustered VMDKs: Round Robin PSP is fully supported. Fixed and MRU PSPs can be used as well, but Round Robbin PSP might provide better performance utilizing all available paths to the storage array.

         Note:

  • While choosing a PSP, consult your storage array vendor for the recommended/supported PSP. For more information, see the Storage/SAN Compatibility Guide.
  • If the number of paths to a storage device exceeds 5, WSFC cluster storage validation may fail if Round Robin path policy (PSP_RR) with iops =1 is used. In such a case, we recommend setting the number of iops to 5 or more when Round Robin path policy is used. This is not a VMware issue.

Perennially reservations

Perennial reservations on the clustered VMDK enabled datastores are required for all ESXi hosts hosting VM nodes with pRDMs and Clustered VMDKs. Check ESXi host takes a long time to start during rescan of RDM LUNs for more details.

In-guest shared storage configuration options.

Maintaining in guest options for storage (such as iSCSI or SMB shares) is up to those implementing the solution and is not visible to ESXi.

VMware fully supports a configuration of WSFC using in-guest iSCSI initiators or in-guest SMB (Server Message Block) protocol, provided that all other configuration meets the documented and supported WSFC configuration. Using this configuration in VMware virtual machines is similar to using it in physical environments. 

Note:

vMotion has not been tested by VMware with any in-guest shared storage configurations.

VM Limitations when hosting a WSFC with shared disk on vSphere. 

Hot changes to virtual machine hardware might disrupts the heartbeat between the WSFC nodes These activities are not supported and shall cause WSFC node failover: 

  • Hot adding memory.
  • Hot adding CPU.
  • Hot adding storage controllers like lsisas, pvscsi or nvme.
  • Hot adding network adapters.
  • Any other hardware changes while cluster VMs are in powered on state except hot add disk or hot share disk.
  • Using snapshots.
  • Cloning a VM
  • Pausing and/or resuming the virtual machine state.
  • Memory over-commitment leading to ESXi swapping or VM memory ballooning.
  • Sharing disks between virtual machines without a clustering solution may lead to data corruptions.
  • Online Extension of a clustered VMDK/vVol is not supported. The kb will be updated when the support for this operation will be provided.
  • Offline Extension of a clustered VMDK/vVol is supported. This operation requires all VM nodes to be powered off. Check the kb https://kb.vmware.com/s/article/67098 for the step-by-step instruction how to extend a size of a clustered VMDK
  • vSCSI filters and I/O filters are not supported with clustered VMDK OR passthrough RDM.

A shared disk resource provided by pRDM can be extended online or offline.

Scsi controller bus sharing should be set to "physical" for SRDF (Symmetrix Remote Data Facility) kind of configurations.

Note:
For more information on VM configuration limitations, see the vSphere WSFC Setup Limitation section in the vSphere Resource Management Guide .

Microsoft support policies for a virtualized deployment of WSFC

Microsoft support deployment of WSFC on a VM. Also check the Microsoft KB article Support policy for Microsoft SQL Server products that are running in a hardware virtualization environment.



Additional Information

Disclaimer: VMware is not responsible for the reliability of any data, opinions, advice, or statements made on third-party websites. Inclusion of such links does not imply that VMware endorses, recommends, or accepts any responsibility for the content of such sites.
 

 

Powered by Wolken Software