Authors: @Devendra Mahajan @SANDEEP BANGUR @Sayali Joshi
The IBM Storage Virtualize Plugin for vSphere 2.1 now supports vCenter storage partitions to manage the entire storage workload for a vSphere deployment. Traditional storage systems often fall short when it comes to handling diverse workloads, maintaining isolation, and keeping costs under control. Storage partitions change that narrative by enabling true multi-tenancy, secure workload isolation, high availability, and optimized costs, all while streamlining management.
Note: For more information on the features of the previous version, IBM Storage Virtualize Plugin for vSphere 2.0, please visit blog
Highlights of IBM Storage Virtualize Plugin for vSphere 2.1
- Migrate Storage Systems from Legacy to Partition mode
- Support for HA partition - PBHA
- Flashsystem Grid awareness
- Manage Storage Systems with vCenter storage partitions
- Manage Host and Host clusters with vCenter storage partitions
- Automated scheduled reconciliation
- Improved datastore performance monitoring and compatibility with latest vCenter 9.0
Migrate Storage Systems from Legacy to Partition mode:
vSphere plugin supports three modes during registering a Storage System:
- Legacy: Registering a storage system to the plugin in the traditional way, without partitions.
- Local vCenter storage partitions (Non-HA): These are partitions created and managed within a single IBM Storage System. They help organize and control storage resources locally, without high availability across multiple systems.
- High Availability (HA) vCenter storage partitions: These partitions span across two systems to deliver a highly available storage environment. Even if one system goes down, the storage remains accessible, ensuring continuous availability and reliability.
Note: vSphere plugin only supports vCenter storage partitions created on storage system
If the user has already registered storage systems in the plugin using legacy mode and wants to migrate the entire legacy workload into the vCenter storage partition, they must perform the seeding process on the storage system. Please refer to the following vlog for detailed steps.
IBM Storage FlashSystem:vSphere Plugin-Migrating existing storage configuration to vCenter partition
Once the seeding process on the storage system is complete and all objects are placed into a vCenter storage partition, the next step is simple: assign the vCenter’s vc_url or vc_guid to that partition. This association enables the vSphere plugin to recognize it. From there, the user just needs to click the Reconcile ALL button in the vSphere plugin GUI. The plugin will then automatically detect the partition and begin managing its objects according to the chosen partition mode Local or HA
Flashsystem Grid view in vSphere Plugin
Plugin provides enhanced visibility and management capabilities for FlashSystem Grid environments. Here's how it works:
Grid Detection & View Activation:
If a FlashSystem storage system is added via the Add Storage flow and is part of a FlashSystem Grid, the plugin automatically enables the FlashSystem Grid View.
This view appears as a submenu under the Storage Systems tab, labelled with the FlashSystem Grid Name.
Grid Overview:
Users can view all systems that are part of the grid. The interface displays total number of partitions across all systems and aggregate storage capacity of the grid.
Unregistered Systems:
If a system in the grid is not yet registered with the plugin, its status is shown as Unregistered. An ADD button is available next to such systems, allowing users to quickly register them into the plugin.
With the vSphere plugin, users gets the flexibility to Add, Edit, or Remove storage systems through plugin and seamlessly associate them with either Local or HA vCenter partitions. When adding a storage system, the plugin first checks if there’s an existing matching vCenter partition. If a match is found, whether it’s a local or HA partition plugin will detect it, and the system will be seamlessly registered with the plugin. If no matching partition is detected, the user has two options: either link the system to an available partition or skip the selection and continue using the storage system without assigning it to a partition.
If a storage system was originally registered in the plugin under legacy mode, and vCenter partitions are later created on the backend storage system, users can easily transition workloads into a vCenter storage partition. This can be done through the plugin's Edit workflow, where the desired partition can be selected from a dropdown to update the systems in desired partition mode.
Check the vlog below for more information on managing storage systems
IBM Storage FlashSystem: vSphere Plugin - Add Storage System with vCenter Partition
For more details about the Manage Storage Systems in vCenter Partition, visit IBM Documentation
Manage Host and Host Clusters with vCenter storage partitions
IBM vSphere 2.1 plugin offers enhanced functionality for managing hosts with partitions. It includes:
Seamless Partition Association:
Hosts and Host clusters can be seamlessly associated with partitions, ensuring efficient management and organisation. When a storage system is registered with a vCenter partition, it is automatically added to the associated partition, eliminating the need for manual configuration with partition.
Location Selection:
When storage system registered with partition which includes HA, then in while adding host user needs to select Location. Setting a host location helps reduce the latency of I/O operations. Here's how it works:
- When a host location is set, the I/O submitted by the host is sent to the system in the same location, while high availability is established.
- If a host location is not set, the host will normally send I/O operations to the storage partition's preferred system.
- In Simple Terms choosing a location helps your system work more efficiently by reducing delays in data transfer.
- User can edit location in edit host workflow.
Check the vlog below for more information on managing host and host clusters
IBM Storage FlashSystem: vSphere Plugin - Host and Datastore operations with vCenter Partition
For more details about the Manage Host and Host Clusters with vCenter storage partitions visit IBM Documentation
Manage Datastores with vCenter storage partitions
vSphere 2.1 plugin allows users to create datastores in partition mode.
-
Key Requirement: Hosts, host cluster, and storage system must be part of the same partition to create a datastore.
-
HA Type Partition: If user wants to create a datastore which involves HA thenOnly the active management system is visible for selection and Datastore is created on both systems.
-
Disaster Recovery:If a volume group with a DR policy is selected, a copy of the datastore is created on the DR system.
Snapshot and Clone Requirements
- Datastore must not be part of the default volume group to create snapshots and clones.
- Cloning of datastores is blocked for HA type partitions.
Check the vlog below for more information on datastore operations
IBM Storage FlashSystem: vSphere Plugin - Host and Datastore operations with vCenter Partition
For more details about the Manage Datastores with vCenter storage partitions, visit IBM Documentation
View Storage Partitions
The IBM vSphere plugin provides a concise view of registered vCenter partitions and their required information, organised into four cards:
1. Replication Overview
- Visual representation of system connections within a partition
- Displays replication configuration, if applicable
- Indicates "No Replication Configured" for local partitions
2. Storage Component
Displays key storage entities within the registered vCenter partition, including Hosts, Host clusters, Volumes, Volume groups, Offline hosts.
3. Connectivity
- Displays partnership status and IP quorum information
- Provides insight into entity partnerships and their connectivity status
4. Partition Information
Displays detailed partition information, including:
- Replication topology
- Migration status
- Partition IP
- Replication policy
Following image shows storage partition of PBHA + DR configuration
Reconciliation
Reconciliation in plugin is essential to come to terms with changes done in Storage system or at vCenter level. There are couple of reasons for plugin to not knowing the changes in first place.
- Actions not initiated via plugin - For instance; Storage administrator deleted the underlying volume without informing VMware administrator.
- Secondary impacts due to resource actions not under plugin’s remit - For instance; Ownership changes at storage system level or new virtual machine deployed by VMware administrator on a datastore managed by plugin.
Plugin supports three flavours of reconciliation. Each has been designed to cater different requirement and flexibility to remain in sync with Storage array and vCenter.
Global reconciliation
This is the overarching reconciliation where plugin try to assess new as well as existing entities.This is akin to reconcile and build ledger from scratch.It gathers and builds information from vCenter and all registered Storage systems.The entities which are of interest are:
Primary entities:
-
Flashsystems: Attributes of registered storage system like version, FlashSystem Grid association, Storage Insight association, role (AMS or non-AMS in case of HA) are updated as part of this.
-
Pools: Only existing registered pools are validated.To update registered pool inventory, user need to use Edit Storage system workflow.
Secondary entities: Hosts, Host Clusters, Datastores, Virtual machines, Sessions between host and storage system/s, System and volume performance historical data, Flashsystem Grid
To initiate global reconciliation, go to vSphere plugin Dashboard, select the Storage System > Registered Systems tab and then click the “RECONCILE ALL” button.
Individual Storage System Reconciliation
User can initiate reconciliation of vSphere entities related to individual storage system as well. This saves from reconciliation everything and rebuild inventory for that particular storage system only and leaves objects related to other storage system untouched.There is a caveat though if a storage system is in HA relationship then reconciliation of one will lead to automatic reconciliation of other as well.
One more aspect where it differ from global reconciliation that it is focused on discovering new secondary entities rather than validating existing ones.If any changes occurs in existing secondary entities listed above , that will be taken care in respective config workflow.
To initiate individual storage system reconciliation,go to vSphere plugin Dashboard, select the Storage System > Registered Systems tab and locate the desired Storage System in the list.Click the vertical ellipsis (⋮) at the start of the row and from the context menu, click “Reconcile”.
Scheduled Reconciliation
To ensure that inventory remains updated and be in sync with Storage system and vCenter, plugin has the facility of scheduled reconciliation which run every night between 2 AM and 3 AM(localtime).This is akin to initiate individual reconciliation for every storage system automatically.
For more details about reconciliation, visit IBM Documentation
Summary
In short, the IBM Storage Virtualize Plugin for vSphere now empowers users to manage multiple IBM Storage systems directly within vCenter storage partitions through the familiar IBM vSphere GUI. With this enhancement, you can seamlessly perform all key storage management tasks like provisioning VMFS datastores, managing hosts and host clusters, handling volume groups and snapshots, while also gaining the added benefit of a FlashSystem grid view, all from one unified interface.
Reference links:
For any queries feel free to contact:
Rahul Fiske: rahul.fiske@in.ibm.com
Sandeep Bangur: sandeep.bangur@in.ibm.com
Devendra Mahajan: demahaj1@in.ibm.com