Cloud Storage Solutions

IBM maintains a longstanding partnership with VMware, and the IBM Cloud Storage Solutions team is a key design partner for VMware Virtual Volumes (VVOLs).

Hear from Carlos Fuente, an IBM Distinguished Engineer, about the capabilities and benefits of VVOLs together with IBM Storage offerings in this YouTube video.

Get your hosts ready, as the new HAK is in now available! The latest 2.9.0 release improves connection stability when attaching hosts to storage systems in IBM HyperSwap and Hyper-Scale Mobility deployments.

As in the last few releases, HAK is available for Windows, Linux, or AIX hosts. A quick reminder for HAK newbies: it allows your hosts to easily connect to volumes on IBM Spectrum Accelerate Family storage systems (FlashSystem A9000 and A9000R, XIV, Spectrum Accelerate deployable SW) over iSCSI or Fibre Channel.

Version 2.9.0 also supports the 1Q18 milestone software release of FlashSystem A9000 and A9000R, version 12.2.1, as well as the RedHat Linux 6.9 and 7.4 releases.

You can download HAK 2.9.0 from IBM Fix Central. Check out the HAK version 2.9.0 documentation on IBM Knowledge Center, which includes the HAK lifecycle and compatibility matrix, latest release notes, and user guide, which now - by the way - has additional info and references regarding host multipathing with FlashSystem A9000 and A9000R. Also, if Microsoft Windows Server is the OS of your hosts, check the new list of required hotfixes (KBs) page (previously listed in the release notes) for error-free operation of your hosts.

As always, your feedback is always welcome.

The IBM Cloud Storage Solutions team

IBM Spectrum Control Base Edition, version 3.3.0 has landed and with it a new major step forward in IBM cloud integration capabilities. Now, you can use IBM storage systems as persistent storage for containerized applications in Kubernetes clusters. This version brings a new cloud interface, IBM Storage Enabler for Containers (1.0.0), a nifty plug-in that makes a smart use of Spectrum Control Base storage profiles. These profiles are used for automating persistent volume provisioning in Kubernetes environment. Pure magic, right?!

Download IBM Spectrum Control Base Edition (3.3.0) and the Installer for IBM Storage Enabler for Containers (1.0.0) from IBM Fix Central. You can also go to GitHuB to see the IBM's open-source project that IBM Storage Enabler for Containers is based on. As usual, do not forget to check out the IBM Spectrum Control Base Edition (3.3.0) release notes on IBM Knowledge Center.

More information about this exciting offering is available in these blog posts:

• IBM Storage systems go cloud native
• IBM Spectrum Access for IBM Cloud Private
• Making container technology work for persistent microservices

A new version of Spectrum Control Base Edition software (3.2.1) has been released. It adds support for version 8.1 of the IBM Spectrum Virtualize Family (a broad range of IBM storage systems), as well as the IBM Spectrum Virtualize deployable software solution. For further details about this new version, including fixes, check out the IBM Spectrum Control Base Edition 3.2.1 release notes on IBM Knowledge Center at: https://ibm.biz/Bdj8gc.

The last several releases of the OpenStack Cinder project addressed various disaster recovery (DR) scenarios. In Mitaka, replication v2.1 (codename ‘Cheesecake’) was introduced. This admin API was designed to solve a true DR scenario, when an entire backend is failed over to another backend on a secondary site. There are other DR scenarios in which the user wants to have better control over the replication granularity. For this purpose, the Cinder team has decided to provide tenant-facing group APIs, allowing volume grouping for a tenant. This enables volume group replication and failover without affecting the entire backend.

This feature (codename ‘Tiramisu’) was introduced in Pike release, and implemented by IBM Storage Driver for OpenStack.

The IBM Storage Driver for OpenStack supports the following IBM storage systems:

• IBM DS8000 Family
• IBM FlashSystem A9000
• IBM FlashSystem A9000R
• IBM Spectrum Accelerate
• IBM XIV Storage System

The user can choose between a generic group replication and consistency group (CG) replication. For consistency group replication, the driver utilizes the storage capabilities to handle CGs and replicate them to a remote site. On the other hand, in generic group replication, the driver replicates each volume individually. In addition, the user can select the replication type. For example, IBM FlashSystem A9000 and A9000R storage systems support both synchronous and asynchronous replication.The minimum Cinder client version that supports group replication is 3.38.

The following functions were added to the IBM Storage Driver for OpenStack to support group replication:

• Enable replication
• Disable replication
• Failover replication

The following example illustrates configuration of a replicated consistency group.

1. Create sync replicated consistency-group.

• Create a volume type for replication.

#cinder type-create rep-vol-1

• Configure the volume type.

#cinder type-key rep-vol-1 set replication_type='<is> sync' replication_enabled='<is> True'

• Create a group type.

#cinder group-type-create rep-cg-1

• Configure the group type.

#cinder group-type-key rep-cg-1 set group_replication_enabled='<is> True' replication_type='<is> sync' consistent_group_snapshot_enabled='<is> True'

• Create a CG, using existing group type and volume type.

#cinder group-create rep-cg-1 rep-vol-1 --name replicated-cg-1

2. Create a volume and add it to the CG.

• Create a replicated volume.

#cinder create --name vol-1 --volume-type rep-vol-1 1

• Add the volume to the CG.

#cinder group-update --add-volumes 91492ed9-c3cf-4732-a525-60e146510b90 replicated-cg-1

3. Enable replication.

#cinder group-enable-replication replicated-cg-1

4. Disable replication.

#cinder group-disable-replication replicated-cg-1

5. Fail over replicated CG.

#cinder group-failover-replication replicated-cg-1

Let’s face it: even the most resilient, robust, and secure storage sites, located above ground or even underground, could potentially be exposed, in a certain degree, to all kinds of disasters that may temporarily or indefinitely halt the site operation. This could be an extreme weather condition or natural disaster (hurricane, tornado, earthquake, etc.), a prolonged and unplanned interruption of power supply (power grid or power station failure), an accident, act of war, cyberattack, and who knows what else – God forbid…

Enterprises that rely on non-stop, continuous, and unbreakable access to data, as well as on the ability to keep existing data integrity in parallel to on-going update/writing operations at the highest possible standards, must have disaster prevention and recovery mechanisms in place, ready to be used at any point in time.

Enterprise-class storage systems, such as IBM XIV Gen3, IBM FlashSystem A9000, and IBM FlashSystem A9000R, provide advanced site mirroring capabilities, either synchronous or asynchronous, including 3-site mirroring for XIV Gen3, and HyperSwap for FlashSystem A9000 and A9000R. These native underlying replication and high availability (HA) technologies allow these IBM storage systems to serve not only conventional hosts and clusters, but also virtual machines in VMware and Microsoft cloud environments, as well as OpenStack cloud nodes.

Virtual machines duplicates can be deployed at a secondary backup site, together with the backup storage systems to serve those sites.

• In a Microsoft Azure environment, automatic failover to the backup or primary site is natively supported by the storage system, by using Microsoft Azure Site Recovery.
• In a VMware environment, virtual machines can be recovered by using VMware Site Recovery Manager (SRM) together with the IBM Spectrum Accelerate Family Storage Replication Adapter (SRA), which also supports – starting from version 3.0.0 – HyperSwap operations.
• OpenStack high availability configurations are also supported.

For detailed information about employing HyperSwap for VMware and Microsoft environments, refer to the IBM Redpaper: ‘IBM HyperSwap for IBM FlashSystem A9000 and A9000R’. Additional data protection scenarios are described in this IBM Redpaper: ‘IBM FlashSystem A9000 and A9000R Business Continuity Solutions’.

Host connection is a vital part of any storage system deployment procedure. Basically, a host is the physical computer or virtual machine that uses storage resources of the storage system, rather than having its own local storage. Accordingly, the hosts are defined on the storage system itself, allowing for this interaction.

Only hosts that are mapped to storage volumes can access those volumes on the storage system. Also, all host I/O operations must be equitably distributed among the storage interface modules. This workload balance is ensured by the storage administrator, who also monitors and assesses it over time when host traffic patterns change.

However, the host operating systems and network protocols come in different types and vendors. This variety presents a challenge to host administrators who must configure the host properly before it can perform on-going and uninterrupted I/O operations on the storage system. This is where the IBM Storage Host Attachment Kit (HAK) comes in handy.

The HAK for the IBM Spectrum Accelerate family – namely XIV, FlashSystem A9000 and A9000R, as well as the software-defined storage (SDS) Spectrum Accelerate deployable software – provides the host administrator with all the required tools for automatic and simpler host diagnostics, configuration, and attachment to the storage system. In addition, the HAK facilitates the monitoring and management of storage volumes from the host.

The HAK can be run on Linux (RHEL or SLES), Microsoft Windows Server, or IBM AIX hosts, and uses iSCSI or Fibre Channel SANs for connecting to the storage systems. Moreover, you can forgo the host installation altogether and use the HAK as a portable package (disk-on-key version). This allows you to run its utilities from a shared network drive or a portable USB flash drive.

Once the host is attached to storage system, defined on the storage device itself, and have the storage volumes mapped to the host, you can:

• Display information about connected storage arrays and volumes.
• Configure connectivity parameters.
• Run host diagnostics.
• Gather configuration information on the host and create host profile that can then be shared with different IBM internal technical support systems.

In all, the IBM Storage Host Attachment Kit is a nifty small-footprint program, indispensable for fast and smart storage deployment. Provided free of charge, it’s available for download from the IBM Fix Central.

For information about the latest version, check IBM Knowledge Center.

It would be quite accurate to state that virtualization technology today is gaining momentum, allowing companies to lower their total cost of ownership (TCO), reduce their datacenter footprint, ensure faster server provisioning, and more. However, implementing server virtualization without accounting for the storage element could cause problems, such as uneven resource sharing, as well as degradation in performance and data integrity.

Various IBM storage systems provide virtual enterprise storage for virtual servers. Specifically, the IBM storage systems work in concert with the following VMware virtualization platforms, standards, and environments:

• VMware vSphere Hypervisor (ESXi)
• VMware vSphere Web Client (vWC)
• VMware vSphere APIs for Storage Awareness (VASA)
• VMware vSphere Virtual Volumes (VVols)
• VMware vRealize Orchestrator (vRO)
• VMware vRealize Operations Manager (vROPs)

To leverage, manage, and centralize this entire connectivity from a single interface, the IBM Spectrum Control Base Edition (SCBE) software package enables a simplified deployment and more efficient integration of different IBM storage systems (also referred to as storage arrays) and the VMware vCloud suite.

IBM Storage with VMware vSphere Web Client (vWC)

Without SCBE, when VMware administrators run out of disk space and request more space from the storage administrators, the storage increase provisioning may take quite a long time to implement and complete. This occurs mostly because of miscellaneous enterprise “red tape” procedures and processes. Moreover, before provisioning additional storage, storage administrators must fully understand the specific needs of the VMware administrators, then find available space on the attached storage systems, create and map storage volumes, and so on.

The SCBE solution uses an abstraction layer through a storage service and pre-defined sets of storage resources (pools) and capabilities (compression, encryption, etc). The services are prepared by storage administrators and provided for VMware administrators for fast and easy self-provisioning whenever required.

IBM Storage with VMware VASA and virtual volumes (VVols)

In the traditional data storage implementations, storage administrators manage multiple VMFS-based datastores with up to 7 TB in size and without explicit indication of storage capabilities. In addition, all VM snapshot and replication procedures are performed on the ESXi server side, without leveraging much faster hardware-based storage system capabilities.

The virtual volume (VVol) approach eliminates the VM granularity by introducing a single large VVol-based datastore with dynamic definition of storage policies (services) for provisioning storage resources exactly according to application requirements.

IBM Storage with VMware vRealize Orchestrator (vRO)

With the IBM Storage Plug-in for VMware vRealize Orchestrator, VMware administrators can include storage discovery and provisioning into their workflows. The plug-in allows issuing automated workflows with storage volumes that are attached to the vRealize Orchestrator server, eliminating the need for manual storage provisioning by storage administrators.

IBM Storage with VMware vRealize Operations Manager (vROps)

Monitoring IBM storage systems without VMware vROPs is limited to information about their physical objects. When using the IBM Storage Management Pack for vROPs, the storage systems are integrated into global monitoring network at the network operations center (NOC). The systems appear on vROPs dashboards, presenting in detail all relations between storage system objects (disks, ports, volumes etc) and VMware objects (datastores and VMs).

This somewhat belated blog post is meant to clarify the past, present, and future development of the IBM Storage Driver for OpenStack. For over 4 years, ever since the OpenStack Folsom release (when Cinder was still part of Nova), IBM Storage Driver for OpenStack has been released by IBM to support the integration of IBM XIV and IBM DS8000 storage systems with OpenStack cloud environments. In parallel, the IBM SAN Volume Controller, IBM Storwize Family, and IBM FlashSystem V9000 storage systems were already included in the OpenStack community source code. Throughout this period, the IBM XIV and IBM DS8000 storage driver remained proprietary. However, as community standards called for opening the driver source code, a decision to incorporate the driver into the community code has been made.

We decided to take the dive into the open-source code in version 2.0.0 of IBM Storage Driver for OpenStack, aiming to include IBM FlashSystem A9000 and A9000R, IBM XIV, IBM Spectrum Accelerate (deployable software), and IBM DS8000 Family systems in the community source code for the OpenStack Ocata release.

The journey of opening the code has been a unique one, as most drivers start as open, and do not evolve from a closed source code. But the challenge was met, and, as we are actually nearing the next release of version 2.1.0, we can make a full use of countless benefits of any open source software. Among them are:

• Security, with multiple reviewers scanning the code, testing it, finding and fixing problems.
• Dynamic development environment, offering powerful code collaboration platforms, such as Gerrit and testing tools, such as Tox.
• Flexibility, allowing the users modify the software to fit their needs.
• Interoperability, breaking the boundaries of proprietary data formats.
• Robust support system with passionate communities of users and developers, providing excellent forums, mailing lists, wikis, documentation etc.
• Documentation available on the OpenStack website, rather than on IBM Knowledge Center.

Stay tuned for our next blog post when the community-based version 2.1.0 is released.

If you are looking for the driver’s up-to-date compatibility matrix, you can find it here.

The long and arduous road to HyperSwap is over! We are happy to announce a release of a new software version of our blazingly fast FlashSystem A9000 and A9000R. This is version 12.1, which brings HyperSwap to our storage clients. HyperSwap solution is an ingenious way of providing high availability based on active-active pairing of storage systems per volume or per consistency group. Each volume or consistency group pair uses synchronous replication to keep both systems updated at all times. When certain conditions apply, an automatic and completely transparent failover is performed, so that the applications experience no downtime. As soon as the actual failure is recovered, the pair is automatically resynchronized.

For easier integration with VMware, Linux, AIX and Windows platforms, we've upgraded these Cloud Storage Solution products:

• IBM Spectrum Accelerate Family Storage Replication Adapter, version 3.0.0 for combining VMware SRM failover capabilities with state-of-the-art HyperSwap high-availability solution.
• IBM Storage Host Attachment Kit, version 2.8.0 for hassle-free attachment of Linux, Windows Server or AIX hosts to FlashSystem A9000 and A9000R storage systems in HyperSwap relations.
• IBM Spectrum Accelerate Family Provider for Microsoft Windows Volume Shadow Copy Service, version 2.9.0 for fully automatic creation of snapshots on HyperSwapped FlashSystem A9000 and A9000R storage systems.

Upgrade now to unleash the full power of the HyperSwap disaster recovery!

Yossi Siles, a Senior Offering Manager at IBM, published this article on the new software version for our all-flash IBM FlashSystem A9000 and A9000R storage systems with IBM HyperSwap support.

You may ask: version 3.2.0 is already available so soon? Our answer: yes. Although earlier this month we released IBM Spectrum Control Base Edition version 3.1.1, version 3.2.0 is a genuine breakthrough, bringing support for the Microsoft PowerShell automation tool, in addition to our already well-established VMware cloud integration offerings. The new IBM Storage Automation Plug-in for PowerShell provides 'cmdlets' for automated storage management, including provisioning, host mapping, volume expansion, and other storage-related tasks.

For the first time, you can download the plug-in from the Microsoft PowerShell Gallery website at: https://www.powershellgallery.com/packages/SpectrumControlBase-Client. For further details about this new version, including fixes and an experimental Flocker feature (might not be included in future versions), check out the IBM Spectrum Control Base Edition 3.2.0 release notes on IBM Knowledge Center at: https://ibm.biz/Bds7gK.

We are happy to announce that we have released several drivers to support the new version of the OpenStack Cinder project: Ocata. This, however, is not an ordinary release, and here's why: previously, the drivers for IBM SAN Volume Controller, IBM Storwize Family, and IBM FlashSystem V9000 storage systems were the only ones included in the OpenStack community source code. Now, this latest release is a huge step forward aimed to include IBM FlashSystem A9000 and A9000R, IBM XIV, IBM Spectrum Accelerate (deployable software), and IBM DS8000 Family systems in the community source code as well, moving from proprietary code to being fully open-source. So henceforth, the IBM Storage Driver for OpenStack is part of the OpenStack Cinder repository. Accordingly, the driver documentation has also migrated from IBM Knowledge Center to the OpenStack community website: https://docs.openstack.org/ocata/config-reference/block-storage/drivers/ibm-storage-volume-driver.html

This release was developed in accordance with the OpenStack Ocata specifications, emphasizing our continuous commitment for providing IBM storage customers with cutting-edge OpenStack capabilities. In addition to the aforementioned integration in the community source code, this driver version includes enhancements in replication functionality (also known as "cheesecake") of the supported IBM systems, as well as improvements in multipath support, and auto-selection of I/O groups for Storwize storage systems.

As always, we are here to help with deployment and operation of our OpenStack drivers. Your feedback is important to us, and we'll be happy if you let us know what you think about this milestone release.

A new version of Spectrum Control Base Edition software (3.1.1) has been released. It replaces version 3.1.0 and brings support for the IBM FlashSystem A9000 and A9000R (12.0.3) storage capacity feature. In addition, it introduces a full compatibility with the current VMware vSphere Web Client Server requirements. For details about this new version, see the Spectrum Control Base Edition release notes on IBM Knowledge Center at  https://ibm.biz/BdswcP.