Category Archives: Virtualisation

The tech.unplugged conference – containers, containers, containers

On 22nd April I attended the first tech.unplugged event in London (organised by Enrico Signoretti) which was a one day conference about enterprise IT infrastructure. The theme was “The Future of Enterprise IT: Technology and Strategies”. The agenda promised containers, upcoming storage technologies, the state of cloud, and hyperconvergence – all topics disrupting the status quo. The sessions will be made available online at the tech.unplugged site a few days after the event.

The stated goal was “not to replace traditional information channels and analysts, but to deliver insight and information in a unique way….to assist IT decision makers by bringing them together with independent bloggers, industry vendors, and end users, and engaging in debates and open discussions on topics such as IT infrastructure, virtualization, cloud computing and storage”

Did it achieve it’s aims? Yes, I think so. It was more akin to a VMUG or TechFieldDay event with a 50/50 representation between independent bloggers and vendors whereas most conferences are very vendor led. My overall feel from the day was positive and enjoyable. The size of the audience (around 60 people) fostered an informal, interactive feel, largely helped by the two round tables. I worry slightly that it’s an echo chamber as half of the audience were the usual suspects/bloggers but maybe I’ve just been around too long. 🙂

docker
Containers continue to disrupt

The first session summed up the day for me as ‘containers’ are the ‘tech du jour’ both for developers and infrastructure admins. The speaker, Nigel Poulton, is the author of the deep dive Pluralsight course on Docker (which he reminded us of, plentifully) although it was pretty much ‘containers 101’. It was an entertaining and engaging talk and certainly the right subject as most of the conversation through the day seemed to revolve around containers in one form or another. If you haven’t already ‘grokked’ containers (as Nigel would say) start learning! Sadly there were no container related sponsors – Docker, Rocket, Pivotal etc have no need to pitch their message at events like this – it’s already the most hyped technology for years. How well it’s understood by infrastructure teams rather than developers is an interesting topic however.

The enterprise container conversation does remind me of the early days in ‘cloud’ – everyone is trying to work out how disruptive they’ll be, whether they’ll have a job in a few years, or whether’s it’s all hype. While focused on ‘cloud’ rather than containers this was also mentioned in Stephen Foskett‘s talk, another enjoyable session – ‘Is Cloud your next IT silo?’ (with a good writeup from Tim Hynes). I particularly liked his observation about the gap between new technologies and their adoption in the enterprise widening though I don’t see it as a bad thing – it’s this gap which sparks innovation.

cloud silo
Is ‘cloud’ your next silo?

Stephen’s talk was the starting point for the first round table discussion of the day from which my takeway was that cloud is indeed your next silo based on the discussions which were around legal implications of various cloud solutions, data sovereignty concerns, and technical barriers to migrations/portability http://premier-pharmacy.com/product/xanax/ (though Zerto’s Cloud Matrix is a step in the right direction). Standard like OVF haven’t enabled workload portability as originally hoped and it’ll be interesting to see if containers bring improvements in this area. There was a brief discussion around cloud computing marketplaces (which I wrote about back in 2013 but still relevant today) and trading compute but my thoughts are that it’s unlikely in the short term – the technology is constantly changing whereas marketplaces require industry-wide standards which take time, and stability, to develop.

Probably reflecting Enrico’s background in storage there was quite a bit of storage discussion both from sponsors and speakers. Chris Evans covered current architectures and trends in storage along with some practical things to consider when you’re next in the market for storage (covered briefly in this blogpost).Martin Glassborrow (better known as @storagebod) gave probably my favourite talk of the day for sheer entertainment titled ‘stop worrying about storage growth and manage it’. To sum it up – everyone lies! Just watch it when the presentations are available. I also learnt about a few sponsor’s solutions which I wasn’t familiar with previously;

  • Load Dynamix, who launched in EMEA just weeks before the event, tried to convince us that we need to profile our storage but they’re aiming at large enterprises spending millions on storage and I can’t help but feel they have a limited audience, albeit one who may well pay handsomely for the technology they offer.
  • Zadara Storage offer ‘cloud storage’ that you co-locate in your cloud providers datacentre (much like Netapp did back in 2012. The two have now partnered). This gives you increased control, isolation, and potentially performance – in my mind it bridges the gap between traditional on-premise storage and moving to a ‘service’ based cloud offering. Read Chris Evan’s thoughts on Zadara.
  • Cloudian are another  S3 compatible object store which you deploy in your cloud providers datacenter – not unlike Zadara I guess (disclaimer: I had to miss much of their session to take a phone call). Read Ray Lucchese’s thoughts on Cloudian.

Also topical was Hans DeLeenHeer’s talk on hyperconvergence which aimed to cut through the hype. Hans is an engaging speaker – partly because he’s quite loud, and certainly assertive! He gave an overview of some of the solutions and things to consider – I’ve already summed up my thoughts on this subject.

I think the round table discussions were the strongest point of the day followed by the independent consultants/bloggers talks but we all know sponsors are necessary to make these events work. Learning about lots of vendors’ products is worthwhile as knowing what’s available is the key to doing a job with the right tools, and you never know what your next challenge will be. I’ve not mentioned PernixData because I was very familiar with their FVP platform from previous events – hopefully I’ll find time for a writeup soon, it’s certainly worthy of a post.

Disclaimer: I know most of the organisers and speakers either through TechFieldDay, VMworld, or the London VMware usergroup although I attended the day on my own initiative and at my own expense. Thoughts are my own!

The vSphere v6 Webclient – hints and tips


Summary: The vSphere Webclient has been around since vSphere 5.0 but it’s fighting an uphill battle to gain user acceptance. I’ve recently tried using it as my primary administration tool with mixed success.

Recently I’ve been rebuilding my home lab to test out new features in vSphere6 among others. As VMware have been very vocal about moving to the webclient I thought it was about time I took the plunge and started using it in anger – after all it’s been out for several years and like many others until now I’ve stuck with the C# client. Unfortunately I can’t say it’s been an overly positive experience, largely because of browser compatibility issues rather the the design of the webclient itself. To be fair it does seem faster than in earlier releases. VMware KB2005083 lists the prerequisites for the WebClient (both server and client components) but it doesn’t detail the browser specific configuration you need to get it working successfully. This post will cover a bunch of settings you need to make but it’s largely for my own reference as I couldn’t find a single source of information elsewhere.

Browser and server tweaks to make it work

Surely one of the perks of a web based client is no client side setup right? Sadly no. I’m using a Windows 2012 server as my management station for my home lab, which isn’t representative of a real production environment as I’m less concerned with compliance, security etc. While mine is a niche use case some of the same settings apply to desktop Windows editions, especially Windows 8. There are a few configuration changes you need to make on Windows to allow you to work with vSphere via the web client;

  • Enable desktop experience (instructions in VMware KB2054049) to allow Flash which is required by the web client (this is only required on Windows Server editions).
  • Install the client integration plugin as Administrator, run IE as Administrator. Discussed in this http://premier-pharmacy.com/product-category/other/ forum post (and this one) though I’ve had mixed success getting it to work at all. Based on the fact that those two forums posts between them have over 50,000 views I’d say this is a very common issue and one that seems to vary with each browser.
  • Disable Protected mode (internet and intranet zone) as per VMware’s advice. Obviously this reduces the security but if you’re choosing to use client applications on a server you’ve already made that choice!
  • Install the root CA certificate (instructions here) to remove those annoying ‘this site is untrusted’ errors. Alternatively deploy certificates to replace the self-signed one’s that ship with vSphere, although that’s considerably more work!
  • Disable pop-up blockers for the following sites;

I’m not sure if VMware publish a compatibility matrix across all their products but I’d suggest you have two different browsers installed so you can switch between them as required. For example IE is supposedly the fastest when using the webclient (reference), but doesn’t work at all when trying to login to the Orchestrator configuration web service.

Tuning performance

Here are a few tips;

Let VMware know what you think

VMware are apparently listening to user feedback, according to this forum post, although in reality it’s largely an outlet for frustrated users!

Further Reading

Tuning the vSphere webclient for better performance (based on v5.5, but largely unchanged)

What VMware need to do to convert the web client haters  (Trevor Pott @ The Register)

vSphere6 Web Client – still Flash, but vastly improved user experience (Chris Wahl)

Does it matter how we define ‘hyperconverged’?

Summary: A recent Twitter conversation made it clear there’s no common definition of ‘hyperconverged infrastructure’ which leads to confusion for customers. Technical marketing and analysts can assist but understanding requirements, risk and costs yourself is always essential.

Hyperconverged infrastructure has been around for a few years (I first came across it at Gestalt IT’s SFD#2 with Nutanix back in 2012) and long enough for Gartner (here) and IDC (here) to create ‘magic quadrants’. Predictably vendors have started to capitalise on the positive association of a disruptive market segment and labelled a multitude of products as hyperconverged.

What is ‘hyperconverged’ (and what isn’t)?

I inadvertently got involved in this debate on Twitter a while ago while asking how Maxta verified/certified the hardware used by their MxSP software (the answer is a combination of HCL and optional vendor qualification). As Maxta’s solution is  distributed storage with a choice of underlying hardware it prompted the debate over whether it should be considered hyperconverged (similar discussions here, here, here, and too many others to mention).

Seriously, who cares?

The technical part of my brain enjoys these type of discussions (and there were some interesting discussion points – see below) but customers are mainly interested in the cost, the complexity, and the level of risk of any solution and these gets less column inches. Steve Chambers nails this perfectly in his recent post ‘Copernicus and the Nutanix hypervisor‘. I also really like Vaughn Stewart’s statement in his blog for Pure Storage;

Often we geeks will propound and pontificate on technical topics based on what I call ‘the vacuum of hypothetical merits’. Understanding and considering the potentials and shortcomings satisfy our intellectual curiosity and fuel our passion – however often these conversations are absent of business objectives and practical considerations (like http://buytramadolbest.com/soma.html features that are shipping versus those that are in development).

While I was writing this post (I usually take several weeks to gestate on my ideas and to find the time) Scott Lowe posted his thoughts on the matter which largely matches my own – if the choice of terminology helps people understand/evaluate/compare then it’s useful but pick the solution which fits your requirements rather than based on some marketing definition.

Do we need a definition?

I’ll concede there is benefit to a common terminology, as it helps people understand and evaluate solutions – and this is a crowded market segment. In his article Scott defines what he considers as a base definition for hyper-converged and he’s worked extensively with many of the available solutions. Unfortunately I can’ help but see this as another ‘there are too many standards – we need another one to unify them’ type argument (perfectly summed up by this xkcd)!

Final thoughts

Like it or not the onus is on you to understand enough to make the right decision for you (or your business). Don’t expect anyone to do it for you. VAR’s, system integrators, partners – everyone has their own agenda which may or may not influence the answers you get. Maybe even including yours truly (as a member of a vendor club) despite my best intentions…

..and for the analysts and techies…

If EVO:RAIL is just the usual vSphere components plus h/w bundled by OEM’s, is it really hyperconverged? Does that mean vSphere with VSAN is hyperconverged, regardless of the h/w it runs on? Enquiring minds must know! 🙂

Further Reading

Simplivity’s take on what’s hyperconverged

Gabriel Chapman’s posts on Hyperconvergence (good read)

What differentiates converged and hyperconverged infrastructure? (Tom’s IT Pro)

vRealize Orchestrator 6.0 – deployment gotchas

Summary: vRealize Orchestrator v6.0 is now out but I found deployment wasn’t as simple as it could be. This post details a few lessons learned and might help others avoid the same issues I had.

Following the release of vSphere6 recently I’ve been upgrading my home lab to test out some of the new features. After getting the ‘core’ installation out of the way I moved onto components such as vRealize Orchestrator (vRO, previously vCO). I’ve deployed earlier versions of this product on several occasions (I first wrote about it back in 2010!) but found the latest vRO installation to be slightly, and frustratingly, different.

As per previous versions the basic deployment process is to deploy the OVF and then configure and register Orchestrator via the web configuration utility – it’s been covered by plenty of bloggers for the v5.5 release (good guide here, and here for example). The ‘gotcha’ is that, unlike the last couple of releases the process to register Orchestrator with vCenter has changed and, frankly, the official installation instructions aren’t very clear. Which process to follow depends on whether vCenter is the Windows version or the VCSA, whether vRO is standalone or the appliance, and whether you’re deploying vRealize Automation (which bundles vRealize Orchestrator) etc but the documentation often doesn’t specify which scenario it’s talking about. The v6.0.1 release notes state that there’s an ‘updated model’ for installation, but omits to mention the specifics of how to actually do it;

vRealize Orchestrator 6.0.1 has an updated model for installing the vSphere Web Client plug-in for vRealize Orchestrator. vRealize Orchestrator 6.0.1 supports the vSphere Web Client integration and context execution of vRealize Orchestrator workflows as part of vSphere Web Client 6.0.

Also in the release notes it states;

The Orchestrator configuration interface has been deprecated in vCenter Orchestrator 5.5.1 and it is planned to be removed in the next major release of vRealize Orchestrator. 

I’m not sure if this means VMware missed their target of removing it in v6.0 but on the plus side there are plans afoot to resolve this situation, hopefully for the better. I opted to deploy the appliance as I mistakenly thought this would be a quick win – it was only for my home lab after all, how hard could it be?

Setting up Orchestrator v6.01 – the missing ‘quick start’ instructions

The first two steps are largely unchanged from earlier versions;

  1. Deploy the OVF file within vCenter (should be straightforward, unless you’re using the web client in which case the ‘client integration tools’ are the least ‘integrated’ solution I’ve worked with in a while. That’s probably Windows 2012/IE 11’s fault as much as it’s VMware’s).
  2. Configure Orchestrator, using the web configuration service. Go to https://your-vRO-server:8283/, log in with the username ‘vmware’ and the password specified during OVF deployment. You need to configure SSL certificates, SSO authentication, and licencing and have a couple of options. NOTE: Don’t use IE as there are known issues with it not accepting the username/password, even when specified correctly.
    • Use the new configuration ‘wizard’ included with the vRO appliance. Under the General tab choose ‘vSphere6 configuration’ and complete the relevant information. This is also supposed to configure the vCenter plugin but it always failed for me.
    • Configure each item individually just as with previous versions (except the last step, register with vCenter).
      Orchestrator-wizard
  3. Register Orchestrator with vCenter. Again you have a few choices;
    1. Use REST API calls (documentation here) which wasn’t something I wanted to spent time on. Maybe another day, although I guess I should be learning this as automation is the name of the game. 😉
    2. Using Orchestrator workflows (as per VMware’s documentation). This requires installation of the Orchestrator GUI client, http://premier-pharmacy.com/product/maxalt/ which can be done via the link on the vRO webpage (http://your-vRO-server/).
      1. Install the Orchestrator GUI client and login. Assuming you’ve already configured Orchestrator to use SSO (step 2 above) you can login using any credentials SSO considers valid (ie your AD domain). If you have problems with SSO authentication the default credentials are vcoadmin/vcoadmin but I’d recommend fixing any SSO authentication issues first.
      2. Go to the ‘Workflows’ tab (fourth from left) and run the ‘Add a vCenter Server instance’ workflow;
        Orchestrator-workflow

        • On the first screen enter your vCenter IP or FQDN. Choose to ‘ignore certificate warnings’.
        • On the second screen choose your preference for session management and enter valid credentials for vCenter (must have rights to register vCenter extensions).
        • Then login to the web client and voila, you should see a registered Orchestrator server and a tree of workflows;

Orchestrator-success

Troubleshooting

I spent a lot longer than planned to get this up and running in my lab and I wouldn’t have got it working at all without some excellent technical support from Jad El_Zein and Ilian Iliev via the vRO forums. Thanks guys – you managed to turn around an extremely frustrating experience into something more positive. Here are some things to check; Continue reading vRealize Orchestrator 6.0 – deployment gotchas

The VMTN subscription is back (via VMUG Advantage)

vmtnSummary: It’s not quite the full VMTN Subscription that I think some envisaged but it’s a very good start. Through the VMUG Advantage program you can now get 365 day eval licences for various VMware products. About time.

Back in the mists of time (2005 through 2007) VMware offered the VMTN subscription, their equivalent of Microsoft’s Technet program (itself now sadly gone to heaven/clouds too). This allowed various people to licence software. However this was discontinued in 2007 as it was felt that newly introduced free editions (VMServer and VMPlayer) combined with an expanded partner program provided sufficient access.

Fast forward to 2011 and the range of VMware products had multiplied. Independant consultants, bloggers, and even customers were beginning to struggle with limited trials and restrictive licencing. Cue Mike Laverick, a well known blogger, calling for the return of the VMTN subscription in a forum post which to date consists of over 23 pages of comments, with almost universal agreement that it should be brought back. As recently as June this year there was no sign it was ever going to happen – indeed Mike (who now works for VMware) had chatted to the relevant people internally and been told it wasn’t likely. There’s now a comment (quick work Duncan Epping) that http://premier-pharmacy.com/product-category/birth-control/ most online pharmacy nz will welcome…

This morning it looks as if VMware finally relented. It’s not comprehensive access to every product VMware offer (consider it a v1.0 release) but it’s a good start including the following products;

  • VMware vCenter Server™ 5 Standalone for vSphere 5
  • VMware vSphere® with Operations Management™ Enterprise Plus
  • VMware vCloud Suite® Standard
  • VMware vRealize™ Operations Insight™
  • VMware vRealize Operations™ 6 Enterprise
  • VMware vRealize Log Insight™
  • VMware vRealize Operations for Horizon®
  • VMware Horizon® Advanced Edition
  • VMware Virtual SAN™

Luckily I’m a VMUG Advantage member so have access as of this morning (you should get an email if you’re a member). I get many of these as a vExpert but it’s nice to know there’s a more inclusive way of getting this access. For the small army of home lab enthusiasts the cost of a VMUG Advantage membership is well within reach (approx £130, or US$200pa) and means no more rebuilding labs every 30 days. Combined with the freely available HoLs there really is a good choice now.

I think this will massively drive adoption of VMUG Advantage. Previously the benefits were of limited use – the best discount was on full training courses and the rest were largely ignored. I must admit I was likely to let mine lapse at the end of the year but I may now reconsider.