Primary Storage

Author: @Ajinkya Nanavati @Sourav Jagati

Introduction:

The storage provider, implemented via VMware APIs for Storage Awareness (VASA), manages all aspects of Virtual Volumes storage. It integrates with the Storage Monitoring Service (SMS) included with vSphere to facilitate communication with vCenter Server and ESXi hosts.

Figure 1: Block Diagram of IBM FlashSystem with VASA Provider

The integration of IBM Storage Virtualize with the Embedded VASA provider has significantly boosted performance and efficiency by leveraging the resources of the storage controller. This fusion simplifies management processes, eliminating the need for an extra virtual appliance. Furthermore, this integration has enabled seamless support for vVol and vVol replication, perfectly aligning with VMware's vSphere Virtual Volumes (vVols). vVols represent a groundbreaking VMware storage technology, revolutionizing the management of storage for virtual machines. Departing from the traditional LUN-based approach, vVols allow for storage management at the individual VM level, offering unparalleled granularity, control, and efficiency.

Figure 2: IBM Flashsystem Supported Operations

IBM Flash-system Storage integration into VMware vSphere, simplifies operations like VM creation, migration, reconfiguration, and deletion across your virtualized data centre. The provider gives you granular control over VM management tasks as well as capabilities for datastore management and snapshot handling. It also enhances business continuity with built-in disaster recovery operations including test failover, planned failover, forced failover, and re-protect functionality. By consolidating these critical virtual machine and storage operations into one unified platform, IBM VASA Provider streamlines administration, improves efficiency, and reduces complexity across your virtual environments.

We will provide a brief overview of the following topics and explore them in detail in upcoming blog posts.

Getting started with virtual volumes (vVol’s) and VASA Provider:

• vVol (Virtual Volumes) is a VMware storage technology that provides VM-centric storage management, allowing administrators to manage storage at the individual VM level rather than the traditional LUN or volume level.
• vVol replication is a feature that allows for the replication of individual virtual machines or virtual disks from one storage array to another, providing data protection and disaster recovery capabilities using IBM Flashsystem Policy Based Replication.
• By leveraging vVol replication, organizations can achieve more granular and efficient replication, failover, and recovery processes compared to traditional LUN-based replication methods. 

Enable vVol configuration through IBM FlashSystem:

•        The IBM FlashSystem family of storage arrays natively supports vVol and provides a user-friendly GUI for configuring the vVol storage provider and creating vVol datastores.
• Through the FlashSystem GUI, you can enable the vVol feature, specify the protocol endpoints (iSCSI or Fibre Channel), and define the storage pool from which vVol datastores will be provisioned.

Integration of VASA Provider with vCenter:

• To integrate vVol functionality in a vSphere environment, the IBM VASA Provider, which is a service implemented by the IBM Flashsystem, must be registered with the vCenter Server.
• After the initial registration, the VASA Provider should be rescanned periodically to ensure that vCenter has up-to-date information about the storage capabilities.
• The VASA Provider can be removed or unregistered from vCenter in case of permanent Storage site failure or storage no longer needed.

Management of vVol Datastore using vcenter

•        In the vSphere GUI, you can register multiple vVol datastores from the same storage system, allowing you to organize and manage your virtual machine storage more granularly.
•            To create a new vVol datastore, you can leverage the vSphere Web Client or the vSphere Client, specifying the storage container, storage policy, and other configuration options provided by the IBM VASA Provider.
•           Removing a vVol datastore from the vSphere inventory is a straightforward process, but it's essential to ensure that no virtual machines are using the datastore before proceeding with the removal.
•            Expanding the capacity of an existing vVol datastore can be accomplished by adding more storage resources (e.g., disks or storage pools) to the underlying storage container, and then extending the datastore through the vSphere GUI or APIs.

Management of SPBM Policy:

·        VMware administrators can define policies based on virtual machine requirements. The IBM VASA provider assists in allocating storage containers that match these policies.

•             SPBM policies are essential for defining the storage requirements and capabilities for vVols, including replication settings that determine whether a vVol should be replicated.
•            For replicated vVols, a replication policy must be created, defining the replication topology and recovery point objective (RPO), in addition to the provisioning policy.
•           Provisioning policies can be thick or thin, matching the storage container's capabilities.
•            SPBM policies can be easily created, removed, deleted, or edited through the vSphere Web Client or vSphere Client, allowing you to modify storage requirements and replication settings as your needs change over time.

Management of vvol enabled VM’s:

    1)  Provisioning New VMs:

•            When creating a new virtual machine in vSphere, you can provision its virtual disks as vVols by selecting the appropriate vVol datastore and storage policy during the VM creation process.
•            For non-replicated vVols, you can select a storage policy that defines the desired storage capabilities.

    2)   Reconfiguration of VM:

             a.      Policy based reconfiguration:

1. 1. You can reconfigure a VM by changing its SPBM policy, which governs the storage capabilities and requirements of replication settings for its virtual disks.
   2. To change the SPBM policy for a VM, you can use the vSphere Web Client or vSphere Client, select the VM, and choose a different compatible policy from the available options provided by the VASA Provider.
   3. The vSphere GUI allows you to display comprehensive SPBM information for a VM, including the current storage policy, compliance status, and details about the underlying vVol objects and their compliance with the policy rules.

              b.     Reconfigure Vvols in VM :

1. 1.  Add a new virtual disk or expand an existing one for the VM.
   2. The newly provisioned or expanded storage will inherit the VM’s current SPBM policy, ensuring consistent storage capabilities across all its virtual disks.

                c.    Snapshot Management for VMs:

1. 1. vVol snapshots capture the state of individual VMs and their virtual disks, unlike traditional storage snapshots that operate at the LUN or volume level.
   2. The IBM VASA Provider enables a chain of snapshots, allowing for efficient and incremental backups of vVols.
   3. This ensures robust data protection with multiple recovery points, optimizing storage use and enhancing disaster recovery strategies. 

vVol replication configuration through IBM FlashSystem:

       1)      Setting Up vVol Replication:

• •        To configure vVol replication on IBM FlashSystem, you can leverage its built-in, policy-based replication capabilities to replicate data between production and recovery arrays.

        2)    Creating VMs with Replicated vVols:

• •        Users can create virtual machines with replicated vVols by selecting a storage policy that includes both provisioning and replication policies.
  •           The replicated vVols will automatically be duplicated to the designated secondary storage system using policy-based replication, ensuring data protection and recovery in case of failure or disaster.

        3)    Migration of Replicated VMs to Non-Replicated VMs and Vice Versa

• •  In vCenter, you can convert a replicated virtual machine to a non-replicated one by changing its SPBM policy from a replication-enabled policy to a non-replication policy, effectively removing it from the replication process.
  •  Conversely, to convert a non-replicated VM to a replicated one, you can modify its SPBM policy to a replication-enabled policy.

Disaster Recovery Operation:

Four primary operations are involved in VMware's Disaster Recovery process, defined below:

        1)  Test Failover:

• • Users can initiate a Test Failover to verify the replication and failover process without disrupting the production environment, allowing them to validate the failover process and test the recovery site before an actual failover event.

         2)     Planned Migration (Planned Failover):

• • For a Planned Migration, they can orchestrate a controlled failover process, synchronizing the data to the recovery site and enabling a graceful switch over to the replicated environment with minimal downtime.

         3)     Forced Failover:

• • In the event of an unplanned outage or disaster, users can perform a Forced Failover, which initiates an immediate and potentially disruptive failover to the recovery site, ensuring business continuity by bringing the replicated environment online as quickly as possible.

         4)     Re-protect:

• • After a successful migration or failover, user can Reprotect the environment, reversing the replication direction and establishing protection for the newly active site, ensuring that data changes are replicated back to the original site or a new target location.

Following VMware tools can be utilized for conducting Disaster Recovery operations.

         1)     Site Recovery Manager (SRM):

• • SRM users can seamlessly execute a Disaster Recovery Plan  using an intuitive GUI interface, allowing for easy validation of DR configuration and orchestration of DR procedures.

         2)     Powercli:  

• • PowerCLI offers the same functionality through a command-line interface, providing automation capabilities for DR procedures, ideal for users comfortable with scripting and command-based operations.

For more details, please contact

@Carlos Fuente

@Rahul Fiske

@Vikrant Malushte

@Warren Hawkins