You wanted the best and you got the Best! – Nutanix on UCS

Today marks a proud day for Nutanix and our customers. As we further extend our lead in the hyperconverged space, it is now fully supported to deploy the Nutanix platform on Cisco UCS servers. Customers now have an additional hardware option to choose from. The current options are NX on Supermicro hardware, XC on Dell hardware as OEM relation or HX on Lenovo hardware as OEM relationship. Outside of these Nutanix currently offers software-only deployments on Crystal ruggedized hardware, open compute project (OCP) hardware and now Cisco UCS.

KISS-cover

The hottest platform in the world, Nutanix on UCS!

Offering the stability, reliability and performance of the Nutanix platform on Cisco UCS has been a regular request from many of our large customers and partners. Customers no longer have to accept other hardware platforms if they are heavily invested in UCS or deploy half-baked or immature HCI solutions that were previously available on UCS.

 

Nutanix on Cisco UCS

Starting today customers can deploy Nutanix on UCS through a meet our meet in the field process. This allows customers to purchase Cisco UCS servers through their normal channels and maintain their Cisco UCS relationships. The hardware and software will be deployed at the customer’s location using the standard Nutanix procedures. The foundation process has been updated to support UCS hardware.

In this initial phase of UCS support, Nutanix will be supporting the C220 and C240 rack mount servers. There will be two models of the C240 to allow for the use of 2.5″ or 3.5″ drives. Also we support deploying with or without Fabric Interconnects (FI), this allows maximum flexibility. These models and config to order flexibility will cover the vast majority of existing use cases. Nutanix will take first call on all support issues and if determined it’s a hardware issue can open a support case with Cisco for customer via TSAnet. Hardware alerts, we can open CiscoTAC cases via TSAnet for customers.

ntnx-ucs-models

 

When deployed with Fabric Interconnects, the foundation process will auto create the necessary identify pools, service profile templates and templates to allow for the normal automated Nutanix deployment process that has been available for years on other hardware platforms.

 

Misc. Faqs

Here are several more details about the release that I won’t dive into at this time.

  • Hypervisor Support ESXi 6.0/5.5, AHV and Hyper-V
  • Regular and self-encrypting drives supported on C240 with 3.5″ drives
  • Haswell and Broadwell CPU’s supported

 

Read More

Home and lab network upgrade with Ubiquiti gear

Recently one of my lab switches began to fail, since it was the one that did most of the routing in my setup it was time to reevaluate my home networking design. I could just pick up another layer 3 switch, drop it in and continue to do the same thing as I was. But I’m always looking to do things better and my current setup was using gear from multiple vendors. I was using Meraki for my firewall and Access Points (AP), HP was my 1GbE networking and routing and Quanta for 10GbE networking. This setup worked fine, but

I was using Meraki for my firewall and Access Points (AP), HP was my 1GbE networking and routing and Quanta for 10GbE networking. This setup worked fine, but obviously there was many different touch points, I would have loved to replace the HP switch with one from Meraki but they are pretty expensive so that was out of the question. Also, I don’t like paying the yearly licensing costs to Meraki but had been doing for a few years because I really liked the features.

So this led me to take another look at Ubiquiti for networking gear, I have seen lots of others express their happiness with the products after using them. So rather than paying for more Meraki licenses in 6 months, I choose to invest that future money and a little more to replace most of my network with Ubiquiti gear. I ended up replacing everything but the Quanta switch that does 10GbE networking.

The new network now uses the Security Gateway (SG) as my edge firewall and router for all traffic. The SG connects to the new 1GbE network switch with is POE capable so it will power the new AP that was deployed also. I use 1GbE for older lab servers and some IPMI connections and then have a trunked connection to my Quanta switch that newer lab hosts connect to. With this setup I now can control all networking expect the Quanta from the single Ubiquiti controller that I deployed on a Windows VM in the lab.

Slide2

While I’m losing a few features that Meraki offered and I used they are things that I can deal with. It’s only been a short period of time but so far I’m pretty happy with the Ubiquiti products and hope they live up to their high praise.

Lessons Learned

I had never used Ubiquiti gear before so there were a few things that I learned while setting up and fighting through some things in the beginning. The first would be to just go ahead and install the Unfi controller software in a VM or an old laptop that will always be on and connected. Install the controller on your laptop is not a great idea if you are not always home and online. The devices hold their configuration but cannot be changed if the controller is not present. You also cannot access the reporting if the controller is not around.

The AP’s are all POE capable which is nice if you do not have power outlets close by where you want to deploy them. They come with an AC adapter or can be powered by a POE capable network switch like the one I purchased. By default the UBNT switch is set to have all ports POE+ enabled, but when I plugged in the AP it would not power up. I tried different cables and nothing worked till I used the AC adapter. After talking to support I found out that you must change the switch port that it’s connected to from POE+ to 24v passive, not sure why this matters but it did the trick. Seems weird that an all Ubiquiti deployment would not power up the AP’s with default settings.

The last weird thing I encountered was that when using my Macbook the performance was not great. It was not obvious when using a browser or even streaming video, but was very obvious when I would RDP to servers in the lab. It would have lots of pauses when click between tabs and apps in the RDP session. If I would keep a ping running to different IPs in the lab I would see random spikes of latency from 15-300ms and a ping that would drop about every 20-30 packets. What was weird is that if I performed the same operations from a PC it worked flawlessly. So off to search the intertubes and I saw that Mac performance on Ubiquiti has been an on and off problem with different firmware versions. There was a bunch of forum posts about the problem and after reading a bunch of them I saw that people were having good luck with running the previous firmware versions on their AP’s.

So I left the controller and switch on the latest firmware versions but downgraded the AP to 3.4.18 and it fixed the Mac performance issues. Immediately after the older firmware was installed and the AP rebooted I had a completely normal experience when performing the same RDP functions.

It’s only been a few days but after working through the issues I’m now pretty happy with my decision to make the switch to Ubqiuti. Now I wish they offered cost-effective 10GbE switches that I could deploy to replace my Quanta and then the setup would be ideal.

 

Read More

Proud to announce that Architecting EUC Solutions book is now available

It’s been a long journey over the past year, but I’m proud to announce that the Architecting EUC Solutions book is finally available. The book focuses on helping you develop your design for modern EUC solutions. It touches briefly on the strategy and roadmap phases of these projects also. The chapters are created so that each one covers a different topic and they range from all of the EUC solutions, operations, infrastructure and all parts required for a design.

The content in the book is very vendor neutral and is not a blueprint on how to write a VMware Horizon or Citrix XenDesktop design. Instead, it takes an approach of educating architects on what questions to ask and how to evaluate alternatives. Then apply these to the solution of your choice.

Pageflex Persona [document: PRS0000038_00067]

I would like to thank Sean Massey for helping by contributing some content for the book and Kees Baggerman for stepping up as a technical reviewer for the book. I hope that if you read the book, that you enjoy it and it’s able to help you on your design journey. If you are not into EUC but are looking for design related content, you may still find some helpful chapters as there are not that many books on IT architecture.

Read More

Monitoring VMs on Nutanix AHV

When managing virtual machines on a Nutanix Acropolis Hypervisor (AHV) cluster there will be a point when you need to monitor a VM or just out of curiosity. This post is going to focus on explaining what data and charts are available to help admins understand the health and performance of a VMs on AHV.

Much like the managing VMs post in the series the monitoring of VMs will be focused on the VM based view within Prism. I have chosen the table view and a sample is shown in the image below. The table provides a list of the VMs that are presented 10 at a time, you can click through them or use the search field to quickly find the VM you are looking for. The table provides basic details such as VM name, the host the VM is running on and the IP address. It also provides the CPU and memory assigned to each VM and then provides a number of storage related metrics. These stats are all available for each VM running on the cluster. The lower portion of this view provides a number of charts that I will go into next.

monitor vm table

 

Read More

Managing VMs on Nutanix AHV

In this post I will take a quick tour through the process of managing virtual machines running on Nutanix Acropolis Hypervisor (AHV). This will touch on the basics and the most common tasks that admins are performing on VMs that already exist. All VM management tasks will be performed from within Prism the single management interface for Nutanix clusters. The VM view within Prism is where the majority of VM-based tasks are performed. A table based view provides an easy to consume method of finding and managing your virtual machines. When selecting a VM, just below the table will be a group of actions that can be taken. These are highlighted in the image below.

These actions are as follows:

  • Enable NGT – Nutanix guest tools, provides VirtIO drivers, Nutanix VSS provider and other services. Similar to VMware tools for vSphere people.
  • Launch Console – Opens a console session to the VM
  • Power Actions – Power off, on, restart actions
  • Take Snapshot – ad-hoc snapshot of the VM
  • Migrate – Use to live migrate a VM to a different host
  • Pause – pause a running VM
  • Clone – Creates a new VM from the existing VM
  • Update – Change VM settings
  • Delete – Delete a VM

manage vm actions

 

Read More

VMware announces App Volumes 3.0 with healthy dose of new features

It looks like VMware is preparing some new EUC nuggets for everyone to consume soon. Today marks the announcement of App Volumes 3.0 and with other events planned in February there seems to be other news coming also.

1av-new-features

 

AppToggle – This feature introduces the idea of app cloaking or the ability to hide applications within a layer if the user is not entitled to them. This provides the option for organizations to create fewer application layers containing multiple applications rather than a larger amount of layers with single apps. In the briefing it was explained that VMware is not just simply hiding the application from the user, they are preventing individual apps from being layered on at the time of the layering being attached to the OS based on individual user entitlements. This sounds pretty interesting and I look forward to getting some hands-on time with this feature.

AppCature with AppIsolation – This feature is integrating ThinApp with the App Volumes capture process. Through a CLI in your capture VM you can enable the application to be isolated, which creates a ThinApp package as part of the process. This allows for applications to enjoy the benefits of isolation where needed and have it all happen in the same workflow. In the past you were able to separately create ThinApps and then include them in a layer, VMware is streamlining the process here.

AppScaling with Multizones – This feature is about working on improving the recovering and multi-site alternatives that App Volumes can support. In the past I’ve been pretty critical of several VMware products for not having a mature story around DR and Active-Active designs. I won’t know if this is a solid alternative or just a good enough one. What is now possible is the ability to use external file servers as the central point to pull AppStacks from. The external file servers has an association with each vCenter server and allows for flexibility in dealing with multiple sites. The idea allows for designs to use DFSR or NAS devices and take advantage of their replication functions, rather than build a replication engine into App Volumes.

Linux-based management VM – The previous versions of App Volumes required Windows for the install of the management console. This new version moves this to Linux which aligns with what VMware is doing for other products and allows for opening up API’s and the unified console story.

Other misc features to mention

  • Layer merging – the ability to merge multiple AppStacks into a single layer (Stack).
  • CLI – A command line interface has been added, which would primarily be used in the capturing and building of AppStacks.
Read More

Radical VMware EUC business ideas

We are almost a month into 2016 and the stream of 2015 recap and 2016 prediction articles and blog posts have finally stopped. But out of these and some recent thoughts, I came up with a few radical ideas for the VMware EUC business unit.I’m well aware that these have pretty much no chance at coming true, but I thought they would make for some interesting conversations. Who knows maybe the right person will read this and it will all come true.

I’m well aware that these have pretty much no chance at coming true, but I thought they would make for some interesting conversations. Who knows maybe the right person will read this and it will all come true.

 

Create an open alternative 

Divorce VMware EUC products from only working on vSphere as the hypervisor platform. This does not mean that the pair is not a good match, it just means that it could be a great opportunity to open up to other hypervisors. Today this is the approach that Citrix offers by supporting multiple hypervisors and offering their customers the choice of platform to deploy on. This flexibly helps by allowing others to drive innovation and offer different cost alternatives when it comes to licensing. With VMware working on building Project Enzo as the future of EUC, which is being built from the code from the Desktone DaaS product. The original Desktone product supported multiple hypervisors, so the code and support was already there to start with.VMware Workspace can still work closely with VMware on vSphere integrations, but they would be focused on using API’s and should be loosely tied to specific vSphere versions. Today if you deploy the whole horizon suite stack with vSphere and maybe even VSAN. Well if you need to update vSphere for a bug or new feature or want the latest VSAN release a hypervisor upgrade is required.

VMware Workspace can still work closely with VMware on vSphere integrations, but they would be focused on using API’s and should be loosely tied to specific vSphere versions. Today if you deploy the whole horizon suite stack with vSphere and maybe even VSAN. Well if you need to update vSphere for a bug or new feature or want the latest VSAN release a hypervisor upgrade is required. A Horizon upgrade usually forces a hypervisor upgrade usually forces a Horizon upgrade unless it’s just a maintenance release. These updates may also force a database version update if you were lagging in your SQL version already. As the stack continues to grow the cascading dependencies has gotten very long and upgrade one later can force a number of other upgrades that you were not interested in doing.

By offering a more open integration with vSphere this could be reduced, also by support other hypervisors you open up a number of other alternatives. VMware has been focused on trying to build the ultimate EUC product suite offering and giving customers a single vendor to deploy. But the reality is that most customers usually are still deploying at least another vendor for profile management (UEM) and monitoring tools. It’s a long a hard process for a single company to build a perfect product suite offering. This does not mean VMware should stop trying but by offering

It’s a long a hard process for a single company to build a perfect product suite offering. This does not mean VMware should stop trying but by offering a all VMware and an open alternative they can give customers the ultimate choice. By having more freed around the

By having more freed around the hypervisor VMware Workspace could build their next generation management layer in any cloud service. This means that they could host or allow customers to build Enzo management layers in AWS or Azure rather than being stuck with VMware Cloud offerings that are not as widely adopted. Below are the current VMware suite offerings, there are a number of options and you can see as you move up there are a large number of products.

Below are the current VMware suite offerings, there are a number of options and you can see as you move up there are a large number of products.

vmware suites.001

 

The open alternatives would allow customers to seek out the best products in the supporting products and not have to pay for VMware legacy products when they want other alternatives.

open suites.001

 

Spin out EUC Business unit

The idea here is that VMware would spin out the EUC business unit into it’s own private or publicly traded company. This idea would make the first idea a lot more feasible. We will call this new entity VMware Workspaces, nobody better register that domain name before I get the chance.

VMware workspaces would own the Horizon (Apps & VDI), App Volumes, UEM, Mirage, workspace, ThinApp, and AirWatch products. They should probably also own and control the Workstation and Fusion products since they are desktop related also.

The new entity can still work closely with the VMware hypervisor team and other teams to continue to build integration between the core products and the EUC products. But the new VMware workspace company is free of any legacy core data center baggage and roadblocks that keep them from innovating and moving to market faster. I think this new freedom could greatly benefit the existing products, processes and allow for the new entity to advance their solutions faster.

VMware Workspace would also be free to offer integrations with 3rd party solutions and provide greater options for customers to build solutions from as discussed earlier. They would still have an all VMware option, but by offering the open option the product licensing can be reduced.

I also think that a separate EUC entity would be free to innovate without having to conform to legacy VMware products. A good example around this would be EUC monitoring. The existing vROPS monitoring tool is heavily deployed for core VM management, but does not do a great job for EUC monitoring. It also is very polluted when you are forced to load different management plug-ins to support different platforms. By being free to create their own monitoring tool a better product could be released.

In the past, this probably would not even be possible, but in the last 2 years VMware has built a great leadership team in the EUC business unit. Along with acquiring several products in this time period they have also grabbed up some excellent leaders, ones that are better to lead a company rather than a business unit. The executive team for a future VMware Workspace entity could be Sanjay Poonen as CEO, Harry Labana a VP, Noah Wasmer a VP, Shawn Bass CTO. Also with growth of the EUC product sales and AirWatch purchase there is now probably plenty of revenue to support a separate entity.

 

Read More
Page 1 of 5112345...102030...Last »
%d bloggers like this: