IT Blog Awards

I’m pleased to announce this blog is a finalist in the 2018 IT Blog Awards hosted by Cisco in the category of “Best Analysis”. Voting is open to the tech community through January 4th 2019, so if you’ve found this blog useful or insightful at any point (or you’ve followed my advice and your datacentre didn’t catch fire as a result) please can you go to http://cs.co/itblogawards and pop in a vote for “IT Should Just Work” in the first section.

If you don’t want to vote for this blog, that’s fair enough, but I’d still recommend having a look at the voting form if only for the awesome “I’m not a bot” section at the bottom.

Thanks!

CiscoBlogAwards_Finalist_Best Analysis

UK VMUG 2018 Preview

IMG_20181106_121248612_editIn two weeks I’ll be heading up to the annual UK VMware User Group meeting. This year a later VMworld Europe has pushed this conference’s date into Christmas season from it’s usual home in November and, speaking of homes, it’s relocated from the National Motorcycle Museum in Birmingham to the National Space Centre in Leicester. “The Sky’s the Limit”? – not here.

The Keynotes are going to be held in the museum planetarium and I’ve heard rumours that Joe Baguely, VMware CTO, is pretty excited about the possibilities this could bring to his opening address. At the other end the day VMware’s HCI Chief Technologist Duncan Epping is giving the closing keynote so you can expect even more great tech coverage here.

It doesn’t end there though, with over 40 sessions to choose from on the day plus over 30 ecosystem vendors represented the UK VMUG is often referred to as a mini-VMworld. It’s large enough to have a range of technical depth whilst still covering a breadth of topics. Sessions cover the entire VMware stack, from vSAN to VVols, HCI to VDI, vRealize to NSX, VMs to Containers, and on-premises deployments right through to the public cloud. I always return from this event feeling I have learnt something relevant to my daily role, and something new for the future.

VMUG-UserconAny good conference starts the night before and this one is no different with it’s vCurry (held in the venue itself) which is usually accompanied by a fiendish vQuiz. Googling on your phone under the table will only get you so far in this one. The vCommunity (who are partly responsible for sticking a v in front of any noun they can) will be out in force there, and right through the event on Thursday.

On the subject of community, in the afternoon I’m talking alongside vExpert Pro Gareth Edwards as we discuss VMware’s global advocacy program; what it involves, the benefits of achieving the award, and how to go about applying to join the current crop of 161 UK vExperts for the 2019 programme.

Checkout the full agenda for more details of this and the other talks.

imageIf you’re reading this before 13th December 2018- Register at the link below, and I’ll see you in Leicester:

https://www.vmug.com/Attend/VMUG-UserCon/2018/UK-VMUG-UserCon-2018

VMworld 2018 Banner

vSAN Scalable File Services

One of the new developments that caught my eye at VMworld this year was the introduction of file services to the VMware vSAN software-defined storage platform. vSAN already offers VMDK storage to vSphere and the ability to host iSCSI volumes, but this feature will allow NFS and SMB file-shares to be hosted directly on the cluster without the need for a separate Windows Server or NFS provider.

What

Yanbing Lee and Duncan Epping discuss vSAN at VMworld Europe 2018

Yanbing Lee and Duncan Epping discuss vSAN at VMworld Europe 2018

vSAN Scalable File Services is a layer that sits on top of vSAN to provide SMB, NFS (and others in future) file shares. It’s comprised of a vSAN Distributed File System (vDFS) which provides the underlying scalable filesystem by aggregating vSAN objects, a Storage Services Platform which provides resilient file server end points, and a control plane for deployment and management.

File shares are created using the vCenter GUI or via API calls from an automation platform, and the demos at VMworld included all the functionality you’d expect with permissions, quotas and so on.

An interesting point is that all the file shares are integrated into the existing vSAN Storage Policy Based Management, and on a per-share basis. Therefore FTT, encryption,  thin provisioning, and so on can all be defined at a pretty granular level. So if only one of your file-shares has an encryption requirement that’s just a case of setting the policy in a drop down list, or likewise if a particular file-share must be configured to be site-failure resilient across a stretched cluster.

Why

Why would you want to do this? Well, a couple of use cases immediately sprang to mind. Firstly, the small office/ remote office/ branch office scenario. A company wants to host both virtual machines and file services in a compact environment- currently the choice would be to have a NAS plus compute hosts, or possibly go hyper-converged but run a VM within this serving the file data from a VMDK. vSAN file services simplifies this by providing that NFS/SMB provision from within the hypervisor- and this also means that all the benefits of resilience, deduplication, compression, and encryption can be provided to the file services.

The second case was for a SAN replacement- a traditional SAN is basically an expandable cluster of x86 servers loaded with disks running some file+disk management software. vSAN is the same thing, but can also run VM workloads. It would be an interesting price/feature comparison exercise to compare the two methodologies.

When

This offering is currently in Public Beta – details at the bottom of this article. NFS 4.1 with AD Authentication is expected at release, with SMB, OpenLDAP, vSAN Data Protection and other functionality to follow. Obviously this is all subject to change as VMware are still at the Beta stage, and a release date has not yet been confirmed.

Further Information

  • HCI3041BE – VMworld Europe 2018 session: Introducing Scalable File Storage on vSAN with Native File Services (Video and Slides)
  • HCI3728KE – VMworld Europe 2018 session:  Innovating Beyond HCI: How VMware is Driving the Next Data Center Revolution (Video)
  • www.vmware.com/go/vsan-beta – Sign up for the Beta. Phase 2 includes the ability to test vSAN File Services in your own lab environment.

VMworld Europe 2018

Vendor Brief: Rubrik-2018 Edition

Following on from a similar briefing at VMworld 2017 I sat down at this year’s event with a couple of the Rubrik execs to see what’s happened in the last 12 (and a bit) months, and what the future holds for the company.

IMG_20181106_155148405_editLast year Rubrik stood proudly on their flagship product- the “Brik” backup appliance and software, but since then their product range has expanded, primarily with their new lines “Polaris” and “Radar”. I see this as a pivot point in their development- they’ve established a solid foundation (both financially and reputationally) and are now expanding the portfolio and also their headcount, growing from ~350 staff to over 1500 in the past year.

However, as Rubrik VP Jerry Rijnbeek was keen to show, this isn’t a random or extravagant direction, but a journey that Rubrik has been on since it’s inception. Polaris is a data management platform, and Radar builds on that to provide anti-malware capabilities- but they’re both tied firmly into the backup application.

The ingenious bit here is that by sitting in the backup tier, the software “sees” all the data in the organisation, be it originating from legacy servers, a virtualised datacentre, or a cloud application. Even if the backup data is not stored on the Brik but pushed out to alternative storage (perhaps a NAS in the datacentre or commodity cloud storage) the Rubrik software already has enough metadata to be able to provide value.

In the case of Radar, the software establishes what normal behaviour looks like for an environment, it can then spot unusual activity- for example folders rapidly getting encrypted could point to encryption malware running riot on a file server, or a quantity of VMs being deleted could indicate a compromised admin account.  This is when Rubriks traditional powers kick in, allowing for live mounting of snapshots and therefore instant restore from a point before the infection took hold.

I got the impression this was only the first step as well- there were obviously plenty of ideas in the company of what other services they could offer having placed themselves at this particular point in the data chain which everything goes past. I’d definitely recommend keeping a watch here as the potential for useful analytics and automation is massive.

Back on the stand in the Solutions Exchange, in a sequel to last years giveaway of the vSphere Host Deep Dive, Rubrik had a thousand copies of the new Clustering Deep Dive book to give away, with the authors on hand to sign each copy.

Thanks to Jerry, Karl, and Vil for taking the time out of their show schedule to sit down with me and to Frank, Niels, and Duncan for their hard work signing books at the giveaway.

Please read my standard Declaration/Disclaimer and before rushing out to buy anything bear in mind that this article is based on a sales discussion at a trade show rather than a POC or production installation. I wasn’t paid to write this article or offered any payment, although in a spirit of full disclosure I did independently pick up a signed copy of the vSphere 6.7 Clustering DeepDive book being given away on the Rubrik stand and enjoyed an evening out at the annual Rubrik VMworld EU party.

Quick PowerCLI- Getting VM hardware versions

A quick PowerCLI snippet for examining what VM Hardware versions exist in your virtual environment:

Using the “Group-Object” cmdlet we can run up a quick count of all the VMs on each hardware version

Get-VM | Group-Object Version

Count Name                      Group
----- ----                      -----
42    v13                       {VM1,VM2,VM3...}
257   v8                        {VM4,VM5,VM6...}
70    v11                       {VM7,VM8,VM9...}
2     v4                        {VM10,VM11}
5     v10                       {VM12,VM13,VM14...}
2     v9                        {VM15,VM16}
2     v7                        {VM17,VM18}

This can be refined using “Sort-Object” to put the most common hardware version at the top of the list.

Get-VM  | Group-Object Version | Sort-Object Count -Descending
Count Name                      Group
----- ----                      -----
257   v8                        {VM4,VM5,VM6...}
70    v11                       {VM7,VM8,VM9...}
42    v13                       {VM1,VM2,VM3...}
5     v10                       {VM12,VM13,VM14...}
2     v7                        {VM17,VM18}
2     v9                        {VM15,VM16}
2     v4                        {VM10,VM11}

We may only be concerned with VMs that are Powered On, so “Where-Object” can be used to filter the original list.

Get-VM  | Where-Object {$_.PowerState -eq "PoweredOn"} | Group-Object Version | Sort-Object Count -Descending
Count Name                      Group
----- ----                      -----
66    v8                        {VM4,VM5,VM19...}
51    v11                       {VM7,VM8,VM9...}
33    v13                       {VM1,VM21,VM22...}
5     v10                       {VM12,VM13,VM20...}
2     v9                        {VM15,VM16}
1     v4                        {VM10}

This quick snippet can be useful when establishing the range of hardware versions in an environment, or estimating the amount of work involved in updating VM hardware to a modern standard across an estate.