Category Archives: EUC

#SWUKVMUG | VDI Made Easy with Nutanix

Last week, I was asked to present at the South West UK VMUG, alongside Nutanix to present a real world deployment story. I talked at a high level about how using Nutanix in a 6K seat VDI deployment, not only made my life easier as one of the the architects on the solution, but how it also helped my client meet their requirements easily.

I’ve linked to my slides below, for those who are interested.

Big thanks to Michael Poore, Simon Eady, Barry Coombs and Jeremy Bowman for the invite to come and speak, also big thanks to Nutanix for allowing me to take half of their presentation time!

If you would like to view the presentation, then you can download it by clicking the link

 

Nutanix and VMware APIs for Array Integration (VAAI) – Quick Tip

In the second of my series of Quick Tip’s with Nutanix I wanted to cover off  VMware APIs for Array Integration (VAAI).

The Nutanix platform supports VAAI which allows the hypervisor to offload certain tasks to the array. This vSphere feature has been around a while now and is much more efficient as the hypervisor doesn’t need to be the “man in the middle” slowing down certain storage related tasks.

Nutanix support all the VAAI primitives for NAS

  • Full File Clone
  • Fast File Clone
  • Reserve Space
  • Extended Statistics

If you are not aware of what these primitives mean, I’d suggest reading the VMware VAAI Techpaper.

For both the full and fast file clones an  NDFS “fast clone” is done meaning a writable snapshot (using re-direct on write) for each clone is created. Each of these clones has its own block map meaning that chain depth isn’t anything to worry about.

I’ve taken the following from Steven Poitras’s Nutanix Bible

The following will determine whether or not VAAI will be used for specific scenarios:

  • Clone VM with Snapshot > VAAI will NOT be used
  • Clone VM without Snapshot which is Powered Off –> VAAI WILL be used
  • Clone VM to a different Datastore/Container –> VAAI will NOT be used
  • Clone VM which is Powered On –> VAAI will NOT be used

These scenarios apply to VMware View:

  • View Full Clone (Template with Snapshot) –> VAAI will NOT be used
  • View Full Clone (Template w/o Snapshot) –> VAAI WILL be used
  • View Linked Clone (VCAI) –> VAAI WILL be used

What I haven’t seen being made clear in any documentation thus far (and I’m not saying it isnt there, I’m simply saying I havent seen it!), is that VAAI WILL only work when the source and destination resides in the same container. This means consideration needs to be given as to the placement of ‘Master’ VDI images or with automated workloads from vCD or vCAC.

For example, if I have two containers on my Nutanix Cluster (Master Images and Desktops) with my master image residing in the master images container, yet I want to deploy desktops to the Desktops container VAAI will NOT be used.

I don’t see this as an issue, however more of a ‘Gotcha’ which needs to be considered at the design stage.

Nutanix Networking – Quick Tip

I’ve spent the past 4 months on a fast paced VDI project built upon Nutanix infrastructure, hence the number of posts on this technology recently. The project is now drawing to a close and moving from ‘Project’ status to ‘BAU’. As this transition takes place, I’m tidying up notes and updating documentation. From this, you may see a few blog posts  with some quick tips around Nutanix specifically with VMware vSphere architecture.

As you may or may not know, a Nutanix block ships with up to 4 nodes. The nodes are stand alone it terms of components and share only the dual power supplies in each block. Each node comes with a total of 5 network ports, as shown in the picture below.

Back_of_Nutanix

Image courtesy of Nutanix

The IPMI port is a 10/100 ethernet network port for lights out management.

There are two 2 x 1GigE Ports and 2 x 10GigE ports. Both the 1GigE and 10GigE ports can be added to Virtual Standard Switches or a Virtual Distributed Switches in VMware. From what I have seen people tend to add the 10GigE NICs to a vSwitch (of either flavour) and configure them in an Active/Active fashion with the 2 x 1GigE ports remaining unused.

This seems to be resilient, however I discovered (whilst reading documentation, not through hardware failure) that the 2 x 10GigE ports actually reside on the same physical card, so this could be considered a single point of failure. To work around this single point of failure, I would suggest incorporating the 2 x 1GigE network ports into your vSwitch and leave them in Standby.

With this configuration, if the 10GigE card were to fail, the 1GigE cards would become active and you would not be impacted by VMware HA restarting machines in the on the remaining nodes in the cluster (Admission Control dependant) .

Yes, performance may well be impacted, however I’d strongly suggest  alarms and monitoring be configured to scream if this were to happen. I would rather manually place a host into maintenance mode and evict my workloads in a controlled manner rather than have them restarted.

Managing a ‘Master’ VDI Image across multiple sites | Made easy with Nutanix

You are working on a large virtual desktop deployment using Active/Active datacenters, you have multiple use cases and multiple master images. With an Active/Active setup, your users have the possibility of being in DC1 one day, and DC2 the next.

So, what do you do when you have a requirement for the image to be available in case of a site failure? Nutanix make this easy for us, using protection domains and per-VM backups.

What is a protection domain?

A protection domain is a VM or group of VMs that can be backed up locally on a cluster or replicated on the same schedule to one or more clusters. Protection domains can then be associated with remote sites.

It is worth noting that protection domain names must be unique across sites and a VM can only reside in one protection domain.

A protection domain on a cluster will be in one of two modes:

  • Active – Manages live VMs, makes, replicates and expires snapshots
  • Inactive – Receives snapshots from a remote cluster

A Protection Domain manages replication via a Consistency Group.

What is a consistency group?

A Consistency Group is a subset of the VMs within the Protection Domain. All VMs within a Consistency Group will be snapshotted in a crash-consistent manner and have snapshots created at each replication interval.

What is a snapshot?

A snapshot is a read-only copy of the state and data of a VM at a point in time. Snapshots for a VM are crash consistent. This means that the VMDK on disk images are consistent with a single point in time. The snapshot represents the on disk data as if the VM crashed. These snapshots are not however application consistent meaning the application data is not quiesced at the time of the snapshot. With some server workloads this could cause us some issues for recovery, however for our VDI master image this is not an issue – the master image is likely going to be powered off the majority of the time. Snapshots are copied asynchronously from one cluster to another.

What are per VM Backups?

A per VM backup give the ability to designate certain VMs for backup to a different site, such as a group of desktop master images. Not all legacy storage vendors offer the ability to replicate at a VM level, normally an entire LUN or Volume replicated at a single time.

 Where am I going with this?

There are many solutions to replicate data, however Nutanix provides this capability, albeit at a small cost, within its platform. No additional components are necessary and it even has an SRM plugin. The key feature is Nutanix integrates with vSphere to make this is a seamless process.

Andre Leibovici posted a great article Disaster Recovery – Failover and Failback with Nutanix which explains the process and includes a video which is below.

Working with vSphere 5.5 and Nutanix  check out my Nutanix Configuration with vSphere 5.5 post

Nutanix Configuration with vSphere 5.5

Over the past few weeks, I’ve been involved in a fast paced VDI project that is planned to scale up to 10K seats.

nosan
Of  late, I’ve not had much involvement with VDI projects and have been focussing more on Private Cloud projects, however this project quickly got my attention as Nutanix were the chosen vendor for the solution.
Given this constraint, the design process was easy, made even easier with the first phase use case.
For those not already aware of Nutanix, here is a great video explaining How Nutanix Works and gives a great insight into the offering.
I wanted to share some of the configuration changes we made to during the build phase and to the vSphere 5.5 platform.

Nutanix Build

First off, Nutanix have recently changed the way they ship their blocks to site. They used to be shipped with a specific flavour of VMware ESXi, however with added support for KVM and Hyper-V, as well as there being a number of different ESXi versions used in the workplace, they found customers always wanting to change the out of the box software. Nutanix now installs the Nutanix Operating System (NOS) controller virtual machine (CVM) and a KVM hypervisor at the factory before shipping. If you want to use a different hypervisor (such as VMware ESXi) nodes must be re-imaged on site and the NOS CVM needs to be reinstalled. Sound daunting? Well, it’s not really.
Nutanix will provide you with two tools named Orchestrator (Node imaging tool) and Pheonix (Nutanix Installer ISO). Once you have these files and your chosen hypervisor ISO, you’ll need to download and install Oracle VM VirtualBox to get underway. This process is very well documented so i’m not going to replay that here, however I would suggest:
  1. Ensure the laptop/desktop you are configuring the Nutanix block from has IPv6 enabled. IPv6 is used for the initial cluster initialization process.
  2. If configuring multiple Nutanix blocks, only image 4 nodes at a time. We attempted 8 at a time, however imaging this many nodes at a time proved troublesome for the installer.
  3. Keep your ESXi VMkernels and CVM VM’s on the same subnet. This is well documented, however for security we had to attempt to split these on different VLANs. This caused some issues with the auto-pathing feature.

I’ll point out here, after we configured out first, 2 block, 8 node cluster, we decided to manually install the second 2 block 8 node cluster and skip using the automated imaging process. This process again is very well documented and took less than 1 hour to have 8 VMware ESXi nodes up and running with storage presented to all nodes. Compare that to the traditional way and this is still a very impressive setup time.

When you’ve built your nodes, boot them into the BIOS and configure the Power Technology from Energy Efficient to Max Performance to ensure power savings don’t dampen performance.

vSphere Configuration

In this particular environment, we were making use of Nutanix de-duplication, which increases the overhead on each CVM. We therefore increased the RAM on each CVM from the default of 16GB to 32GB and set a vSphere reservation to ensure it always has this physical RAM available.

Nutanix have done a very good job with their documentation which detail recommended configurations per vendor for things such as HA and DRS in vSphere. After reading an Example Architecture blog by Josh Odgers I decided to add an advanced parameter to my HA configuration to change the default isolation address by adding “das.isolationaddress1 | Nutanix Cluster IP address”. I chose the cluster IP address over a CVM IP address for a simple reason. If a CVM hosting the cluster IP address fails, the cluster IP automatically is moved to another CVM in the cluster. The cluster IP is a new configuration option that was released for Hyper-V support, but we can make good use of it in the VMware world.

Each CVM resides on local SSD storage in the form of a 60GB SATA DOM. When you logon to the vSphere client and try and deploy a new workload you will have the option of deploying out to this 60GB SATA DOM SSD storage. This deployment was solely going to be used for a VDI project, therefore all workloads would be provisioned directly from a broker meaning we can control in the broker which datastores the workloads will reside in.  So, to avoid any confusion and to stop an over eager admin deploying out to the SATA DOM disk, I created a Datastore Cluster with all Storage DRS and IO options disabled and named it along the lines of “Nutanix NOS Datastore – Do Not Use”.

Overall, the Nutanix devices are very easy to deploy and you can be up and running in next to no time. Now, I’ve managed to get all this down in a post, I can do some performance tests for my next post!

Calling all Thin Client Vendors

Aside from my day job, buying a new house, studying for VCAP exams, presenting at VMUG’s (The Belgian VMUG is next week), I have a couple of personal side projects I’m working on, both virtualization related of course. The first exciting project is well under way and I hope to announce it very soon. They second is still in it’s infancy and is the reason for this blog post, so I’m hoping it works!

I’d like to reach out to all thin client vendors who offer  thin clients with Teradici PCoIP chips built in to ‘donate’ or ‘loan’  a unit to me for some testing. You may be thinking, I work for a well known virtualization consultancy, surely I have some contacts I could call upon. Yes, I do, however I want to be completely upfront and honest this probably wont lead to a sale and is not in anyway associated with Xtravirt. You may also be thinking why don’t you just go and buy one for yourself. I could, but I’ve just bought a house, also buying a single unit wont get me any relationship with a vendor. Whilst I’m not intending to pester any vendor for information, perhaps some questions may come from my testing.

So why provide me with a unit? Well, I’ll certainly blog about your kindness, so please ping me a message on Twitter and let’s have a chat!

 

VMUG Belgium | I’m presenting!!

The Belgian VMUG team will be celebrating their 20th VMUG meeting on December 5th 2013 at the San Marco Village. They have pulled together some very big names from the VMware community including Chad Sakac of EMC, Joe Baguley of VMware and Dheeraj Pandey of Nutanix, as well as the technical wizardry of, to name just a few of Lee Dilworth, Cormac Hogan, Luc Deekens and Viktor van den Berg 

I’m very honoured and pleased to say I will also be amongst the speakers at this event. I will be speaking about a recent 4K seat VDI project I was heavily involved in. My main goal of the session is to highlight some of the pain points of deploying VDI at this scale, whilst feeding back some key points to help VDI deployments such as this succeed. This session is an interactive one, so if you plan on attending, please come prepared to get involved!

The meeting agenda for the day is below.

Meeting Agenda

Registration for the event is now open, if you haven’t registered already, what are you waiting for? Click Here!

Look forward to meeting plenty of new faces on the day.

 

 

Citrix PVS | Windows 7 BSOD

During a recent project, I was at the deployment stage of the PoC. This particular piece of work was using VMware vSphere as the underlying hypervisor, with Citrix XenDesktop as the broker solution and Citrix Provisioning server for the image deployment.

It was a relatively small PoC, just to test out the designs to ensure they would stand up in the pilot phase, and to ensure we had relevant communication ports open.

The install went very smoothly, until I started to create the VHD’s with Provisioning server and boot test machines from that VHD image. Whilst the image conversion completed without any ‘major’ errors in the log (there were a number, but not to worry about) and the image we copied was in perfect health, the VHD would constantly produce a BSOD.

Now, I had already followed Citrix recommendation to ensure the servers and the desktops were making use of the VMXNET3 network adapter. I had also installed the Microsoft hotfix described here to resolve the known issues with the VMXNET drivers to no avail.

After going through a few different troubleshooting steps mentioned in forums (such as removing the MS hotfix above, then removing the PVS software from the target, reinstalling the hotfix the the PVS software) I was starting to become a little frustrated. A colleague then cast his eye over everything when we discovered the the VMware tools were out of date on both the desktop image and the PVS server. Although these had been initially updated after the VM templates were deployed  the hosting infrastructure was upgraded from 5.0U2 to 5.1 U1 during that day, carried out by a separate team within the customers IT department. As soon as the tools were updated to the latest revision on both the PVS server and the desktop image, the image was re-imported into the PVS server, and as if by magic the issue was resolved.

Lesson learnt – Keep an eye on VMware tools versions and ensure (where possible) they are up to date and consistent when using Citrix PVS.

My Study… What’s Next?

I’ve decided, after a week break from study to pursue my next exam milestone(s), in the form of the VCAP exams for Data Center Virtualisation. It seems to make logical sense to focus on the Data Center design exam first, having just spent a lot of time studying for the desktop design exam. I’ve spent the past 18 months, more hands off the day to day administration of vSphere and focusing predominately on design work so, the VCAP-DCA will take a bit more effort wiping away the cobwebs, therefore I’m aiming to complete in the early part of next year.

I plan on completing study guides covering off the objectives in the same way I did for the VCAP-DTD. I appreciate many have done this before me, and there are too many sites to mention here, however I find this way of studying the most beneficial, therefore will continue, If I can help others in the process, all the better!

Before this however, and partly due to a change in direction in my work role to focus more towards private cloud, I took advantage of a recent discount offer and booked my VCP-IaaS exam, which I’ll study for alongside my VCAP-DCD studies. I’ve decided not to focus as much time writing study guides for the VCP exam, however decided I’d share a list of materials I’ll be using to focus on the VCP-IaaS.

  • The exam blueprint, the first port of call for any exam. I’ll download and review the blueprint, then refer back to this throughout my study to ensure I can confidentially cover off all the requirements in each objective.  I’ll also read all the suggested official documentation in the blueprint.
  • My home lab. You need hands on exposure on all exams. I’ll predominately be using the vCD appliance (which I already have setup) however I’ll install vCD in a RHEL server a few times, to ensure I can cover this off.
  • TrainSignal (or PluralSight if you prefer) offer three vCloud focussed training courses:
    • VMware vCloud Director 5.1 Essentials by Chris Wahl
    • VMware vCloud Director Essentials by David Davis and Jake Robinson
    • VMware vCloud Director Organisations by Jake Robinson

As the first title suggests, Chris Wahl’s series is based on 5.1, whereas the exam focuses on 1.5. There is still bucket loads of good information in here, so well worth watching. If you do not subscribe to TrainSignal, it truly is worth it, subscribe here!

If I use any further materials, then I’ll update this post. The VCAP-DCD study guides should also start coming fairly soon!

VCAP-DTD | What materials I used to study

As part of my recent studies towards the VMware VCAP-DTD I used various training materials, so, I thought I would cobble a post together letting everyone know what I used, and what I thought was best.

Before anyone asks, no, I didn’t use an exam cram, that’s what makes the VMware VCAP exams different to many other industry qualifications, it’s one you can be proud of achieving through dedication and hard work as you can’t cheat!

I was lucky enough to attend the two day instructor led View design workshop to kick off my studies, however I think this course can really be hit or miss depending on the students in attendance. Unfortunately, I found myself in a fairly quiet group, so most of the banter I looked forward to, discussing other peoples design decisions didn’t happen and at times I found myself and the instructor having one to one design discussions with the rest of the group looking on and sometimes looking a bit lost. Don’t get me wrong, this is not a dig toward the instructor, in fact I thought he was actually very good, if it wasn’t for his persistence with the remainder of the students it would have been a boring couple of days with little takeaway. Overall, I’d recommend the class, however just hope you get an active group if you are planning on attending. Even though the course wasn’t spectacular discussion wise, the course material was good and very comprehensive and ended up played a major part in my study.

After the course, with the exam booked (with plenty of time to study further) I downloaded the blueprint and used this as the base of my study. I can’t recommend highly enough to know the exam blueprint inside out. Everything asked of you in the exam is covered off in the blueprint. Read all the suggested material as everything is in there. Head over to Jason Langer’s site, he has kindly downloaded all the material and bundled together in a single zip file, even putting all the files into their respective objective folders. Good work Jas, thanks!!!

After reading through all the official VMware material, I re-visited some of the books I had in my collection around VMware View.  I’ve mentioned these before, they can be found here.

At this point, I started my own study guides, which of course can be found here. Where I took each objective from the blueprint and wrote study notes on each element to ensure I was going to be comfortable with every section of the blueprint. The info in my study guides is a mixture of real world experience and information obtained from all the above sources, but hopefully written in a way which is simple to understand and to digest. Hopefully it will assist many of you looking to sit the exam. They are not intended to suggest that the official courses are not worthwhile, or the books are not worth reading, far from it. I’ve just cobbled together my study notes and posted them in the hope they help others as they helped me pass the tough, but passable VCAP-DTD.

I also signed up to Trainsignal. At $49 a month with no contract tie-in, or $39 a month when you sign-up for a year, I think these are simply fantastic value for money. Whilst perhaps the material they have doesn’t have direct relevance to the VCAP-DTD exam, they do have a series on View, although more aimed at the VCP-DT, they have a series on VMware vSphere design by the one and only Scott Lowe which is simply brilliant and is a fantastic aid in this exam. I’ve started some of the other courses on their site recently to assist with other studies and I cannot recommend them highly enough. With their recent acquisition by Pluralsight, you now have a whole heap more content available at the same price as before!! If you are not already a member I’d seriously recommend you take a look here

I also wrote a post on the exam experience which can be found here