By Simon Stevens, Dell ISG Engineering Technologist, EMEA. December 2024
At the time of penning this blog, it is not unusual for one’s thoughts and focus to turn towards the upcoming holiday season. Traditionally, this is the time of year when both Change Control Boards and the weather in the Northern Hemisphere align, ensuring that the word “freeze” has more than one possible meaning! The holiday season traditionally gives those responsible for maintaining datacentre infrastructure some much needed “downtime of the acceptable variety”. Being able to relax, even for a few short days, is always welcome; it is doubly welcome to technical teams who have very likely spent a lot of the past twelve months “keeping the lights on” for their respective businesses and/or customers. Christmas truly is a time for giving thanks to all of those who add magic into our lives.
A backdrop of IT Operations teams “trying to do more with less” dovetails nicely into the benefits of Dell PowerFlex – the ultimate software-defined infrastructure platform that, among other things, reduces the total cost of operations by 40% while also providing dramatically improved business application performance and way less unplanned business downtime. But it is not just me who is saying this – it is our customers! IDC recently interviewed a varied mix of PowerFlex customers about their experiences. Their feedback was used to create an updated “Business Value Snapshot”, a paper which summarises the business value that customers gain when using PowerFlex. These benefits include seeing improved application performance, obtaining greater staff efficiency across the board, along with the almost complete elimination of unplanned downtime. More highlights are shown below:
The full IDC Business Value Snapshot for PowerFlex 2024 can be found here.
In keeping with the initial theme of the holiday season being the time for giving, a raft of improvements and updates have also been made to PowerFlex and its ecosystem recently. So, sit down and buckle up for the ride, as I take you through a whirlwind tour of the recent updates to the world of Dell PowerFlex at a speed that is faster than any sleigh ride…
PowerFlex M&O 4.6.1 Release
I am starting this tour by visiting what is probably the most important recent announcement. PowerFlex M&O 4.6.1 saw its GA release on 27th November 2024. This is the first “in-market” update to PowerFlex 4.6. As well as an including an updated version of the PowerFlex Management Platform, this release includes an update to the “core” PowerFlex software as well – version 4.5.3 of PowerFlex Core, to be exact. Customers with support contracts can access these latest versions from the Dell Support Site. For software-only/custom node customers, the latest releases can be found by selecting the ‘4.6.1’ radio button under the PowerFlex Software section of the Dell support site.
What are the main features included in this release? For me, this version includes several major “asks” that I know will please existing PowerFlex customers. The main features that I am going to call out here are as follows:
- Ability to Deploy PowerFlex Manager Platform into Linux KVM Environments for appliance customers.
- Automated Broadcom ESXi Upgrades via PowerFlex Manager Automation for appliance and rack customers
- Support added for new Operating Systems and Hypervisors
Linux KVM for PowerFlex Manager Platform on PowerFlex appliance
This feature is one that many PowerFlex appliance customers have been waiting for! With PowerFlex 4.6.1, it is now possible to deploy the PowerFlex Manager Platform (PFMP) onto a Linux KVM hypervisor for PowerFlex appliance customers. It must be noted that for now, such deployments will have to follow a manual process that is provided by Dell. Likewise, customers will be responsible for the setup of the Linux KVM environment that PFMP is deployed into. With 4.6.1 then, we are providing our customers with the QCOW2 images for PFMP, along with the manual deployment process. This release allows the deployment of PFMP as well as Cloudlink for appliance customers that use KVM without first needing to obtain an approved RPQ for this. Expect to see automation for such deployments to come with a future release of PowerFlex (watch this space!).
Broadcom ESXi Upgrade via PowerFlex Manager Automation (for appliance & rack systems)
Again, this is an important feature that many customers have been waiting for. This release includes the automation that facilitates the upgrades from ESXi 7.x to 8.x, as well as upgrading from earlier versions of ESXi 8.x to newer versions. For PowerFlex appliance customers, you are looking for the “46.381.00” Intelligent Catalogue bundle if you wish to use ESXi 8.0, or “46.376.00” if you are using ESXi 7.0. (For PowerFlex rack customers, the equivalent RCMs are 3_8_1_0 and 3_7_6_0).
It is also worth me highlighting that due to recent changes in Broadcom licensing terms, the Intelligent Catalogues do not contain any ESXi images anymore. Customers must supply these themselves by downloading them from the Broadcom Support Site. The standard ISO and Upgrade depot ZIP files can be used and PowerFlex Manager will then allow customers to upload their supported Broadcom images into PFXM and be used for subsequent activities on Resource Groups that use these VMware images.
Support Added for New Operating Systems and Hypervisors
As always, it is important to ensure that PowerFlex keeps up to date with what is happening in the wider infrastructure market. Market demands must be listened to and then responded to in a timely manner if we are to remain relevant in the minds of our customers. Hence the core PowerFlex 4.5.3 software has added support for (or updated support for) the following operating systems:
- Rocky Linux 8.10 & 9.4
- Alma Linux 8.10 & 9.4
- Debian 12.5
- Ubuntu 24.04 LTS
- Windows Server 2025
- Red Hat Enterprise Linux 8.10, 9.3 and 9.4 (updates)
- SUSE SLES 15 SP6 (Software-Only Support)
With the addition of full PowerFlex component support for Debian 12.5, this also means that Proxmox VE version 8.2.x is now supported. We have had customers running Proxmox with PowerFlex for nearly 18 months using NVMe/TCP connectivity, and now we have full support for Proxmox using the SDC. Proxmox customers that decide to opt to use the SDC will need to use the “Self-Compiling SDC” functionality (this is a topic that I must come back to soon in a separate blog). But again, this is great news for customers who are considering their hypervisor options and this adds another “tool to the toolbox”, so to speak.
As well as the features highlighted above, PowerFlex 4.6.1 also includes the ability to migrate existing SVMs that are still using embedded CentOS to embedded SLES and, as one would expect, it also includes a number of fixes to keep on improving the overall PowerFlex experience.
Dell Container Storage Modules Version 1.12
It was only a few weeks ago that the latest quarterly update of the Dell CSM – Container Storage Modules was released – you can see the full changelog for this release here. This version has added a key new feature, support for KubeVirt, which also then provides support with OpenShift Virtualisation. Clearly, one of the major strengths of PowerFlex is its support for a wide range of operating systems, Kubernetes distributions and hypervisors. Customers who presently rely on large estates stuffed full of Virtual Machines but are looking to adopt Kubernetes often hit a range of barriers that prevent adoption of Kubernetes. Having the ability to run VMs that cannot easily be containerized is one way to overcome a lot of these barriers, and KubeVirt allows this to happen – see here to find out more about KubeVirt.
Hence, CSM adding support for KubeVirt is important. The latest CSI drivers for PowerFlex have been tested and validated to do the following:
- Create Virtual Machines backed by Dell Storage
- Clone Virtual Machines from a template
- Take snapshots of Virtual Machines
- Live migrate the virtual machines
- Import Virtual Machines using Openshift Migration Toolkit for Virtualization
The other major changes with CSM 1.12 are the addition of support for Kubernetes 1.31 and OpenShift 4.17, along with the squashing of a number of bugs. Full details of this update can be found In the latest release notes for the Dell PowerFlex CSI driver on Github here. Remember folks that Dell provides the Container Storage Modules/CSI Drivers for free, alongside full support through the usual Dell Support channels for customers with existing support contracts.
APEX Block Storage for Public Cloud – Terraform Modules for Deployment
Hot off the press (and just in time for Christmas!!) Dell APEX Block Storage customers can now utilise Dell-provided Terraform modules to deploy PowerFlex 4.6 onwards onto either AWS or Azure. This is great news for both customers and partners who wish to simplify the deployment of APEX Block Storage into any of the currently supported AWS and Azure regions. These new modules have been added to v1.2.0 of the Terraform Modules for PowerFlex, which were released on the 6th December 2024. The Terraform modules were created to package multiple resources together in a way that makes sense when it comes to performing specific tasks. Previous versions of these modules had already added the capability to install the SDC packages onto different host systems, specifically Linux, ESXi or Windows. There is also a user module that can be used to create, update the roles and delete PowerFlex users.
Being able to automate the Deployment of APEX Block Storage using Terraform will certainly help customers when standing up their PowerFlex environments off prem. In order to use the Deployment Modules, the Terraform Provider for PowerFlex must also be installed, so let me provide you with update on where we are with this. Currently at version 1.7.0, the Dell PowerFlex Terraform Provider can be used to manage SDCs, volumes, snapshots, snapshot-policies, storage pools, SDSs, protection domains, devices, users, MDM cluster, fault sets, firmware repository, peer systems, replication consistency groups, replication pairs, NVMe hosts and NVMe targets. It works with all currently supported versions of PowerFlex (that is, 3.6.x, 4.5.x and 4.6.x) and Terraform 1.8.x and 1.9.x.
The latest version of the “Terraform Provider for PowerFlex” can be found here on registry.terraform.io/providers/dell/powerflex/latest, while the latest version of the “Terraform Modules for PowerFlex” can be found here at registry.terraform.io/modules/dell/modules/powerflex/latest. My colleague Robert Sonders has recorded a great video that runs through an APEX Block Storage for Azure deployment using the Terraform Modules, that can be accessed from the Dell APEX YouTube playlist here. As always, please check out all of the prerequisites prior to deploying these modules or the Terraform Provider itself.
More Open Ecosystem Options for PowerFlex with Canonical
Before bringing this update to a close, I also need to highlight some recent announcements – that both Canonical LXD and Canonical Openstack are now supported with PowerFlex. 2024 has certainly been a year where customers have been dealing with a lot of uncertainty regarding the infrastructure software stack vendors that they continue to have business relationships with. It has been a year where nearly every customer that I have spoken to has been considering their options when it comes to future infrastructure and hypervisor stacks. Of course, Canonical are a major player in the Open Source world and we have long had many PowerFlex customers who have chosen to use their products. As such, we recently announced that Dell PowerFlex is now supported with both Canonical LXD and also Canonical OpenStack.
While Canonical LXD had previously added support for PowerFlex using NVMe/TCP connectivity, version 6.1 of LXD now extends things by adding support for the Powerflex SDC as well to access Powerflex Storage Pools. As PowerFlex is a supported Storage Driver under LXD, it means that PowerFlex is integrated into the LXD UI, making it easier to operate an LXD cluster with PowerFlex storage. A whitepaper jointly authored by Dell & Canonical can be found via this link here on the Dell PowerFlex Infohub. This WP discusses the deployment of an LXD Cluster into a two-layer PowerFlex configuration.
The White Paper first walks through the steps needed to deploy the LXD cluster and then shows how easy it is to setup PowerFlex as a shared storage resource in LXD. It then ends by taking the reader through an example of creating an LXD VM instance that uses PowerFlex storage. You can find more details about this solution by visiting this blog written by my colleagues Tony Foster and Vittal Sanadi here: Canonical LXD and Dell PowerFlex – You Don’t Know Until You Try It .
Turning our attention now away from LXD and onto Canonical OpenStack. For those of you who have been on another planet for the past 15 years, OpenStack is a free, open-source software cloud computing platform that has steadily and stealthily built up a large customer base over the years. Companies tend to gravitate towards OpenStack when they wish to build flexible, scalable cloud platforms that are not tied to one specific vendor. Built around numerous component modules that all started life as individual OpenStack community projects, of particular importance to PowerFlex customers are the OpenStack Cinder & Nova modules. Nova is the component that provisions compute resources, while Cinder is the service that provisions block storage resources.
Support for Dell PowerFlex was up-streamed directly into the OpenStack Cinder driver with the Liberty release back in 2015. Being part of the native Cinder driver makes it easier for customers to deploy PowerFlex block storage into clouds that use OpenStack.
Many open source software vendors have been providing & supporting their own OpenStack distributions over the years, with Canonical being one of the leaders in this space. Canonical’s Charmed OpenStack uses Juju charms to simplify the deployment and lifecycle management of OpenStack applications and resources (a charm is a pre-packaged operator that simplifies certain operational tasks).
The Dell DevTech developers have created PowerFlex Cinder and Nova charms to help make it super simple for OpenStack users to use Dell PowerFlex storage when building out scalable, automated cloud infrastructure. Both of these charms can be accessed from Github – The PowerFlex Cinder charm can be found on here and the PowerFlex Nova charm can be found here. You can find complete details on how to configure and use these charms in this brilliant blog written by my colleagues Prasar Kodati and JP Roquesalane here: Simplified Deployment and Lifecycle Management of the Dell PowerFlex Cinder Driver on Canonical OpenStack.
In Conclusion…
I hope that all of the “updates” that I have highlighted above prove that there is a lot happening within the world of Dell PowerFlex and its ecosystem. That rate of innovation shows no signs of slowing down during 2025 and I am looking forwards to penning new blogs that contain exciting news of yet more new releases over the coming months. I will finish this blog by wishing you and your loved ones all the best for the coming holiday season. But if you are reading this blog in the middle of 2025…. well, at least I got my festive message across to you early this year!
Great to see some real choice coming back now with Proxmox, Canonical & Kubevirt!