The following changes are introduced in SFW to support fast failover:
New dynamic disk group state Deported Read-Only
SFW supports two disk group states; Imported and Deported.
When storage resources are online on a node (the active node), the disk group is imported in a Read/Write mode. In this state the disk group is accessible and writable on the active node. When storage resources are offline on a node (the passive node), the disk group is in a deported state. In this state the disk group is not accessible from the passive node.
With fast failover, a new state, Deported Read-Only, is added to the SFW dynamic disk group entity. The disk group status on passive nodes now displays as Deported Read-Only.
In this state, the disk group is imported in a "Read-Only" mode on the passive node. All the volumes are in an offline state and no transactions are allowed on the disk group.
During failover, SFW normally performs a complete disk group deport operation on the active node followed by a Read/Write import operation on a passive node.
With fast failover, instead of performing deport and import operations, SFW now performs only a mode change for the disk group. The disk group state on the passive node is changed from Read-Only to Read/Write.
A mode change (Read-Only to Read/Write) is a much faster operation as compared to a full deport and import (Deport None to Import Read/Write) and thus results in faster disk group failovers.
The following table describes the different disk group states. The 4th column shows the new Deported, Read-Only state introduced as part of fast failover.
Changes to vxdg command
The SFW vxdg command is modified to display the new disk group state introduced for fast failover. The vxdg list and vxdg dginfo options now display the new Deported Read-Only state of fast failover enabled disk groups.
Here's an example of the vxdg list command displaying the new state:
C:\>vxdg list Disk groups are.... Name : dg1 Disk Group Type : Dynamic (Cluster) DiskGroup ID : 689ef763-9c1f-4e96-807e-1b97543b3dba Status : Deported Access : Read-Only
Here's an example of the vxdg dginfo command displaying the new state:
C:\>vxdg -gSQL_DG dginfo Disk group information for the disk group SQL_DG Name : SQL_DG DiskGroup ID : b1f9e838-0689-4b91-b737-cd33f60cdcf4 Disk Group Type : Dynamic (Cluster) Private Group Protection : No Status : Deported Access : Read-Only Current Version : Yes Version Number : 60 Comment : Number of Volumes : 5 Number of Disks : 1 Names of the Volumes are.... \Device\HarddiskDmVolumes\SQL_DG\REGREP \Device\HarddiskDmVolumes\SQL_DG\DB_LOG \Device\HarddiskDmVolumes\SQL_DG\DB_VOL \Device\HarddiskDmVolumes\SQL_DG\DATA_FILES \Device\HarddiskDmVolumes\SQL_DG\SRL1 Names of the Disk are.... Harddisk2
The vxdg command has a new option, vxdg refreshff, which refreshes the state of disk groups on a system to support the new Deported Read-Only state. You must run this command if the disk group state on a passive node does not display the Read-Only state even though FastFailover is enabled for the VMDg resources. This typically occurs in cases where a new node is added to a service group where FastFailover is already enabled.