Skip to main content
Pure Technical Services

Step 04 -- Disk Policy Configuration (SAN Policy)

Currently viewing public documentation. Please login to access the full scope of documentation.

Disk Policy (aka SAN Policy)

The new Disk Policy (previously known as the SAN Policy) is the policy Windows Server uses to determine whether or not disks should automatically mount that are detected as new on the host. 

An issue has been identified in Purity//FA which cases Windows Server volumes to go offline after a Purity//FA upgrade. Following a Purity//FA Upgrade, if a Windows 2008 R2, 2008 R2 (Service Pack 1), 2012, 2012 R2 or 2016 server hosting FlashArray volumes is rebooted, Windows Server takes the volumes offline. This was due to Purity//FA returning the current version of Purity//FA in the “Product Revision” field in response to an SCSI INQ sent from the host.

The upgrade would change this "Product Revision" leading Windows Server to flag the disks associated with that volume as offline after reboot, requiring Administrator interaction before restoring connectivity, as a safety measure. The Admin would need to access Microsoft Disk Management to manually right click on each impacted disk and manually bring them back online. Purity//FA versions greater than 4.7.10 (4.8.7 4.10, 5.0) now returns 8888 as the current version into the “Product Revision” field, avoiding the behavior of constantly changing version numbers. Consequentially, if a disk is not set to onlineall in the SAN Policy it will be set to offline after the next reboot. Future upgrades of Purity//FA will no longer change Product Revision from 8888 thus fixing the issue, additionally, a script has been created to set the registry values for all SAN policies to onlineall after a server reboot. please see Windows Reporting Drives Offline After Reboots.

This issue has been fixed in Purity//FA 4.7.10, 4.8.7 and 4.9.3.

Windows Server defaults to OfflineShared which is the correct setting when using Windows Server Failover Clustering. In the case of using Windows Server as a standalone host the disk policy should be set to OnlineAll to bring existing volumes online when the host restarts.

If the Disk Policy was changed after having Pure Storage FlashArry volumes connected to a Windows Server host the new disk policy will not be applied to the previous connected volumes. The only way to update previously connected volumes (LUNs) is through the use of Windows PowerShell or DiskPart. See How to set the Partmgr Attributes registry value using PowerShell for full details. Also see Windows 2008 / 2012 Reporting Drives Offline After Server Reboots.

To check the current disk policy run the following Windows PowerShell.

Get-StorageSetting | Select-Object NewDiskPolicy

PS C:\> Get-StorageSetting | Select-Object NewDiskPolicy


To change this to the recommended setting run the following Windows PowerShell:

Set-StorageSetting -NewDiskPolicy OnlineAll


Policy Setting Effect


 All new disks are left offline by default.


 All disks on busses that are detected as internal are left offline as default.


 All disks on sharable busses, such as iSCSI, FC, or SAS are left offline by default. 

 OnlineAll (Recommended)

 All disks are automatically brought online. 


On Windows Server 2008 R2 or 2008 R2 Service Pack 1 the SAN Policy can be changed using Windows PowerShell with the following command.

"SAN Policy=OnlineAll" | diskpart 

DiskPart can be used directly with the following commands.

Microsoft DiskPart version 10.0.14393.0
Copyright (C) 1999-2013 Microsoft Corporation.
On computer: SERVER01
SAN Policy  : Offline Shared
DiskPart successfully changed the SAN policy for the current operating system.


When going from a pre 4.8.8 Purity to a post 4.8.8 Purity with Clustered Windows Hosts using CSVs. 

CSVs, OfflineShared disk policy, regarding CSVs in a Cluster and a host that did own a volume going offline for a reboot: Cluster Manager Service (Failover Cluster Manager) mgs disk connections. One host will always have ownership of a volume (not all on same host usually). Cluster Mgr does a heartbeat for volumes and when a host owner of a CSV loses its vol access the Mgr will notice and find a new host owner on its own to online the volume, without you having to do it manually.