Nutanix Acropolis and XenDesktop – Operations Made Easy – NPX / VMturbo

    VDI is hopefully ran/maintained by your desktop team, not your highest paid guys on the operations team. One of the main reasons of why the Acropolis hypervisor was created was to control the experience as the desktops are managed and as the environment scales. Two of the most sought out after questions in designing a VDI solution is how many desktops per LUN/NFS/Container and how many desktops per management resource(vCenter/SCVMM). On the number of desktops per container for NFS on vCenter for Nutanix the VPXA agent would have problems around 2,000 desktops per container. This is still a far cry from typical 150 for block storage and 250 desktops for traditional NFS based storage but something you needed to account for. With the Acropolis hypervisor there is no need to keep track of the number of desktops running in a container so it makes that design consideration pretty easy.

    Nothing was worse of me on Saturday afternoon updating my VDI environment waiting for the desktops to update in a serial nature. This is the main reason why both vCenter and SCVMM are architectured to only support ~2,000 desktops for VDI, any more than you’ll be waiting around as you update your desktops. Since Prism (Nutanix Management) is based on scale out principles (Map Reduce/Apache Cassandra) we can parallelize the service requests across the whole cluster. Anything added benefit is that Prism is highly available so you don’t have to worry about a single point of failure in your environment.

    Avoid the bottleneck of virtualization management with Acropolis.

    Avoid the bottleneck of virtualization management with Acropolis.

    The Acropolis hypervisor doesn’t have memory over-commitment today but between Large Pages from Windows 7 on and highly dynamic environments were machines are consistently rebooting/re-created memory over-commitment is not as strong as it once was in the Windows XP days. The Acropolis hypervisor can place machines on boot were the majority of their data resides and based on available resources. To further the integration of data placement and compute resources check out what VMturbo is doing for QofS on Nutanix.

    In simplifying the day 2 support for VDI, user experience will go up with consistent performance and availability. Acropolis is freed from the burden of past virtualization management choices and continue to provide one-click everything to run day to day operations to ease. Looking forward to seeing the Citrix and Nutanix relationship to grow even more over the coming months.


    FSLogix For User Profiles and Application Support For Any Hypervisor

    A couple weeks ago I had a great conversation with Cassondra McAllister who is an SE at FSLogix. She has a lot of field time around user profiles as she was at RTO before they were bought by VMware. Cassondra has been working a FSLogix from the early stages and I was lucky to get a demo from here a couple of weeks ago.

    fslogixFSLogix uses image masking to create a single unified base image that hides everything a logged in user shouldn’t see, providing predictable and real-time access to applications and profiles for VDI/XenApp. Companies can combine all applications, plus browser and app plugins, onto a single gold image, or greatly reduce their current number of images. Combine this with quick clones across Acropolis, hyper-v & vSphere and the fact that FSLogix works in the user space can have a pretty simple solution for VDI and XenApp/RDS solutions on Nutanix. This is all possible because FSLogix works at login on time to add the appropriate VHD to the desktop while maintaining fast logons.

    User profiles with FSLogix have shown shorter logon and logoff times, with an improvement of +50% compared to roaming profiles and use less traffic and processing load on file server(s), since files are only accessed if needed. Combine this with Nutanix’s up and coming file services and you’ll have a strong and easy combination for VDI.

    The core of the FSLogix Apps software agent is the driver (frxdrv.sys). This component is a file system mini-filter driver. It is primarily responsible for intercepting requests from other software on the system, to access objects such as files or registry keys, and changing them. When an application is being hidden, the changes might involve simply making the object appear to not exist.

    The second major component in the FSLogix Apps software agent is the service (frxsvc.exe). This component is a Windows service running as system. It is responsible for communicating various data about
    the system state to the driver, e.g., informing the driver when new users login to the machine.

    The other components of the software agent are various user interfaces and supporting files. There is a GUI rule editor that assists administrators in creating rule sets, hiding applications installed on a system, making rule assignments to control how rules are applied, testing rules, etc. There is also a command-line interface that can be used for scripting many of the functions that the software can perform. A Windows Event Viewer integration module is also provided to assist administrators in the management and audit of the software.

    GUI Rule Editor

    People should really check FSlogiX out for cost effective solution that has low overhead from both required infrastructure and overhead on the desktop.


    Introduction to DevOps with Udacity

    Me and Karl

    Me and Karl

    This course was an awesome side project. I was lucky to help with an Intro to DevOps side by side with Karl Krueger. Karl really is the star of the show. Karl worked at Google as an SRE for over 8 years and has some great insightd from the development side. Most of my life has been on the infrastructure side of the house but I did have a brief journey in development in the early 2000’s. I think both sides of the fence have allowed me to have a lot of empathy to my developer friends.

    Like Karl, all the people at Udacity are top notch and made me question how they can pay for all the talent they have. Gundega Dekena did a lot of work half way across the world too. Gundega did a ton work with the course content drawing on personal experience as well. When you take a Udacity course you really get a team of people producing the content. I would encourage you to check out some of their free content as well and if you like it, grab a subscription. The video work used for demo’s is really amazing stuff from what I have seen on the market.

    Let me know if you take the course and how it went.



    Is Hyperconverged Infrastruce More Secure Than Traditional Infrastructure? #Podcast

    An interesting talk about what Nutanix is doing around Security with Eric Hammersley, a security architect with Nutanix. Eric talks about Security development life cycle, STIGS and using automation to tighten lock down the hatches of the XCP platform. Eric talk about the advantages of knowing your own platform and how that can enable engineering to make the most secure platform.

    I find it funny how the most hated guy in the data center (the security guy) is now getting a lot of respect. I am sure the endless supply of credit card information ending up on the world wide web is helping to fuel this. Give the podcast a listen and let us know what you think.


    Podcast Lollapalooza – EUC Podcast and Frontline Chatter

    Last week I was able to sneak onto two podcasts. End User Computing Podcast and the Frontline Chatter Chat. I have to report with deep sadness that the 2016 .Next Conference is NOT in Sydney, my love of rugby must have had the better of me. The 2016 .Next conference is in Las Vegas.

    EUC Podacst was mostly around Synergy and cloud management with a sprinkle of layering and the Frontline Chatter was great review on the announcements for .Next including what Acropolis is a whole.

    Listen to both today and let me know what you think.




    Journey to a Hybrid Cloud: Up and Down and All Around

    Coming off the heels of .NEXT and the announcement of Application Mobility Fabric (AMF) and thought I should really start to take a look at vCloud Air since it wasn’t a part of our announcement. Apart of AMF is moving workloads to cloud providers and it will abstract a lot of the mess that is involved in moving to different virtual machine formats. Going from a vSphere to vCloud Air, VMware has done a pretty good job at extending out the management. In reality, Nutanix fully supports going from on-prem vSphere to vCloud Air.

    I wanted move to a template from my vSphere management cluster to vCloud air and see how the life cycle management would pan out. So I set off and installed the vCloud Connect Server & node appliances on my private cluster. That was pretty easy as it was just deploying some OVA’s. I even just let the appliances run with DHCP and everything was fine.

    One stumbling I had was setting up the connection to vCloud Air from the vCloud Connect Server. It took me a bit to find my organization ID and as thought I only needed to add in the name of my virtual private data center.

    The name of the virtual private cloud was my red herring.

    The name of the virtual private cloud was my red herring.

    Once I figured out I actually had to go into the “Manage with vCloud Director” portion I found the proper ID I should have been entering in.

    Here I was able to find the organization name for vCloud Connect Server

    Here I was able to find the organization name for vCloud Connect Server configuration

    Maintaining different cloud environments may take some work without proper tooling for dev/test.

    Maintaining different cloud environments may take some work without proper tooling for dev/test.

    So I moved my template over, deployed a VM and then went looking on how I could reconfigure the workload. It was pretty easy and I didn’t have to re-deploy from my private environment. I started to look at the AWS process and it wasn’t as smooth for windows based systems. Installing AWS CLI and python to convert images so you could create an AMI (Amazon Machine Image) wasn’t great but do able. The exporting was also CLI based. I guess you have to have a love for CLI for AWS but that’s probably ok for people really into Dev/Test but something to consider. Depending on how you operate, you’ll have to possible recreate your AMI’s if you need to make changes. It might kinda end up being a game of snakes and letters if your comparing the two public clouds coming from a vSphere environment. There is a pile of information about the vCloud Air service here, pricing, user reviews and use cases.

    Free $300 vCloud  Air Credits

    Free $300 vCloud Air Credits

    VMware’s vCloud Air allows you to work in a hybrid sandbox; no need to bring down the latest production version of applications, configurations, workloads. Simply operate in the vCloud Air sandbox and use the latest updates. Nutanix fully supports vCloud Air and the AMF will be used to flatten some of the nuances of moving between clouds over time. I would encourage you to take a free test trail available on vCloud Air and see if fits your workflow.


    Nutanix Volume API and Containers #dockercon

    When I first heard that Nutanix was creating a Volume API to give access to virtual machines via iSCSI I thought it was just arming our customers another option in running MS Exchange. I was sitting in Acropolis training last week and it dawned on me how great this will be for containers and OpenStack. So what is Nutanix Volume API all about?

    The volumes API exposes back-end NDFS storage to guest operating system, physical hosts, and containers through iSCSI. iSCSI support allows any operating system to use the storage capabilities of NDFS. In this deployment scenario, the operating system works directly with Nutanix storage bypassing any hypervisor.

    Volumes API consists of the following entities:

    Volume group
    iSCSI target and group of disk devices.
    Storage devices in the volume group (displayed as LUNs for the iSCSI target).
    Allowing a specified initiator IQN access to the volume group.

    The following image shows an example of a VM running on Nutanix with its operating system hosted on the Nutanix storage, mounting the volumes directly.


    Now your OpenStack and containers instances can be blown away and your data will persist! I think this is a big plus for Nutanix and containers running on the Acropolis hypervisor. Future integration with Docker Plugins should now be easier.


    Tech Preview of Nutanix File Services

    A quick run through of what is cooking for Nutanix File Services. On Acropolis will be able to hide the file services VM’s in Prism.


    Nutanix PowerShell CMDLETS Reference Poster

    Head on over to get the new Nutanix PowerShell Reference Poster

    nutanix powershell


    Nutanix Tech Preview: Prism Management for Containers

    Below is a video of Prism being able to manage containers either virtual or based in AWS. Watch as containers are deployed on Acropolis and AWS.