In another of the series of Virtual Volume information from our partners, today I am pleased to offer a post by Manuel Lange, product manager for the ETERNUS storage systems at Fujitsu. Thank you Manuel for the material, our customers always appreciate the details on how things work in real implementations!
VMware Virtual Volume (vVols) is a new technique that leverages a set of APIs (VASA based) allowing vSphere to communicate with ETERNUS DX S3 disk storage systems.
Combining leading performance architecture with Automated Quality of Service (Auto QoS) management the Fujitsu Storage ETERNUS DX series aligns storage resources with business priorities, thus enabling high system utilization and consolidation.
The communication with vCenter is handled via the ETERNUS VASA Provider which is installed on a server that runs the ETERNUS SF management suite.
ETERNUS SF storage management software supports overall daily operations within the entire ETERNUS DX storage infrastructure by offering an easy-to-use graphical user interface for all management tasks.
Each Virtual Machine that is created using vVols functionality is treated as an own entity and stored in Flexible Tier Pools (FTRPs) on the storage array.
Virtual Machines are stored in a Storage Container that is a logical storage construct placed on top of Flexible Tier Pools with one, two or three layers (Sub Pools). The Storage Container capacity is based on the physical storage capacity of this Flexible Tier Pool.
For any data connection the IO flow is handled through Protocol Endpoints (PEs). A PE enables the communication between the vSphere host and the ETERNUS DX S3 storage array. ETERNUS DX supports PEs using either Fibre Channel or iSCSI technology. VMware uses the PE as a message box to perform IO operations to the ETERNUS DX storage array. From a technical point of view a PE is comparable to a LUN but has no data storage behind it.
ETERNUS storage capabilities
Storage capabilities are array based features that can be used to classify VMware storage policies. The capability defines which features and functions the storage array can offer to the storage container. These capabilities can then be used in one or more storage policies to define VM requirements.
The following five storage capabilities are supported by the ETERNUS DX S3.
Virtual Machine Backup
This capability provides a granular definition of the VMs snapshot schedule. These snapshots can be used to backup and restore an individual VM using array based snapshot technology (SnapOPC+). In addition to the array based snapshot the creation of an array based clone (QuickOPC) to protect against raid group failures is supported as well.
Automated Storage Tiering (AST)
This capability allows defining the desired automated Storage Tiering (AST) settings for each vVol. It allows specification of the number of tier layers as well as the definition of the desired quota shares for each vVol. Quota shares can define how a vVol should be distributed across the different tier levels e.g. 10% on Tier1, 20% on Tier2, 70% on Tier3.
Automated Quality of Service (Auto QoS)
This capability ensures that particular applications always get a certain, predefined performance level. Adjusting the priority and defining a response time enables automatic tuning of the I/O performance and bandwidth to make sure the required response time per application will be achieved.
This capability allows using the Extreme Cache Feature (with flash storage located directly in the controller enclosure or SSDs located in the disk enclosure), if installed. This enables the VMware administrator to provide higher performance while using fewer disk drives.
This capability specifies whether controller based disk encryption should be checked or not (128-bit AES technology or Fujitsu Original Encryption).
These granular capabilities enable the vSphere administrator to actively change the storage behavior to meet requirements. The storage administrator defines the physical resources that can be used for virtual volumes but the management of resources is up to the vSphere administrator.
The deeply integrated automated storage tiering and automated quality of service capabilities can be mapped to the storage policies to granularly define the different VM performance SLAs.
Virtual machine backup capabilities leverage the arrays snapshot and clone capabilities to create an automatic VM aware backup.
The following section illustrates the vVols functionality integrated in the central management suite called ETERNUS SF.
First of all a list of access paths is shown. An access path is the connection between ETERNUS DX storage and the ESX servers. The screenshot below shows two paths across two different VM HBAs connected via two different Fibre Channel fabrics to an ETERNUS DX storage array. Creating an access path includes automatically creating the necessary PEs.
Now that the PEs have been set up, the next operations - like creating a vVols datastore and creating VMs in it - can be performed.
The following screen shows the list of vVols on a particular vVols datastore and how they are mapped to the physical pool as well as the Virtual Machines.
As described in the capability section, a policy based backup schedule for each virtual machine can be created.
This screenshot shows which kind of backup schedule is configured for each individual virtual machine.
In addition the ETERNUS vVols implementation provides two extended restore options. The Single Item Restore functionality allows administrators to mount a vVol snapshot to the source VM. This allows the recovery of individual files from that vVol.
The additional clone backup capability described above protects against raid failures and provides an additional level of protection compared to the already existing snapshots that can help against logical failures.
This demonstration video provides a closer look:
For further information on Fujitsu’s vVols implementation check the following website.