Category Archives: PernixData

Top Posts 2016

2016 is pretty much done and dusted and it’s been an good year for Virtualization is Life! There was a more modest 70% increase in site visits this year compared to 2015 and a 2600% increase in visits since I began blogging in 2012. In 2016 I managed to produce 124 posts (including this one) which was slightly up on the 110 I produced in 2015 and in doing so passed 300 total blogs since I started here. I was fairly consistent in getting out at least eight blogs per month with June being my most prolific month with sixteen blog posts published.

Looking back through the statistics generate via JetPack, I’ve listed the Top 10 Blog Posts from the last 12 months. This year the opinion pieces seemed to be of interest to my readers and there is still vCloud Director and NSX representation in the top ten with my Veeam articles doing well. Again it was interesting to see that two of the most generic (older posts) and certainly basic posts took out two of the top three spots. It shows that bloggers should not be afraid of blogging around simple topics as there is an audience that will appreciate the content and get value out of the post.

  1. NSX Edge vs vShield Edge: Part 1 – Feature and Performance Matrix
  2. Quick Post: E1000 vs VMXNET3
  3. vSphere 6.0 vCenter Server Appliance: Upgrading from 5.x
  4. ESXi Bugs – VMware Can’t Keep Letting This Happen!
  5. Nutanix Buying PernixData: My Critical Analysis
  6. New NSX License Tier Thoughts and Transformers
  7. CBT Bugs – VMware Can’t Keep Letting This Happen!
  8. Veeam 9 Released: Top New Features
  9. Veeam’s Next Big Thing – Veeam has Arrived!
  10. vCloud Director 8: New Features And A New UI Addition…

I was honoured to have this blog voted #44 in the TopvBlog2016 and even with all the controversy around the voting I still hold that as a significant outcome of which I am very proud and I’d like to thank the readers and supporters of this blog for voting for me! And thanks must also go to my site sponsors who are all listed on the right hand side of this page.

With me moving across to vendor land it’s going to be interesting to see if I can keep up the variety of posts as I “narrow” down my core focus…however I fully intend to keep on pushing this blog by keeping it strong to it’s roots of vCloud Director and core VMware technologies like NSX and vSAN. I have the Home lab and the drive to continue to produce content around the things I am passionate about…and that includes all things hosting and cloud now with a touch of availability 🙂

Stay tuned for an even bigger 2017!


Nutanix Buying PernixData: My Critical Analysis

Overnight The Register posted an article claiming that Nutanix is about to buyout PernixData…this has apparently come through reliable sources and hasn’t been denied by those who have been asked at PernixData. I tweeted that I was pretty bummed at the news and as a PernixData Customer and PernixPro I feel inclined to comment on why I do feel that PernixData missed out on being able to maybe go at it alone. This isn’t going to be a post around Nutanix or why they felt they needed PernixData technology, though it’s apparent that they potentially required a performance boost somewhere along the line in their architecture. Possibly they where after the advanced analytics that Architect provides…or maybe they needed both.

FVP is Brilliant Tech!:

There is no doubt that FVP is brilliant and pretty much anybody who has it deployed will attest to the fact that it delivers as promised. On a personal note it came to our rescue when we had performance issues in one of our storage platforms and allowed us to deliver services with low latency and decent performance. The ease and elegance of the solution meant that you could install FVP within 15 minutes across an existing Cluster utilizing investment in flash or memory meant that it should have been a no brainer for a lot of people with existing performance issues.

Apart from the “band-aid” use case the premise of in host caching should have lead storage and platform architects to consider installing FVP into hosts for accelerated performance of read and write IO and allow for cheaper dumb storage to provide the capacity. It’s something that I guess serves as the basis for more storage platforms but the beauty here was that you could potentially cater the solution to fit specific requirements or budgets and have the flexibility to upgrade/downgrade when required.

The introduction of using memory to cache was truly amazing and the possibilities for extreme caching though FVP does have some limitations and gotchya’s but overall it’s very slick technology.

Architect is Very Handy:

Architect was released last year and is installed with the FVP binaries so it’s implanted into the kernel ready for action. Once unlocked it presents probably the best set of platform analytics in the industry. The granularity of the data and metrics it presents is brilliant for any architect or operational person to use in either planning for, or managing storage. The potential to combine FVP and Architect into a full self healing analytics platform was pretty exciting. The Register post mentioned the fact that PernixData where looking at developing their own storage platform to combine all elements…this would have…and still may make sense though there is a lot of competition in the storage array space and maybe the risk of getting into the hardware game means this was just a rumour.

The Problem:

While I don’t pretend to fully understand the ins and outs of how a company prices their solutions, but I can say from experience that FVP and Architect are expensive. I’ve been involved in trying to justify spend on FVP internally and I can tell you that it was/is a hard sell. For the most part if there wasn’t a need to solve a storage problem most would, and have, found FVP too expensive. This isn’t to say that it doesn’t add value, cause certainly it does, however when you break it down and look at the cost of FVP compared to a physical storage array the numbers look kind of sick…especially given that FVP is just software that utilizes existing or new hardware.

Based on some internal workings, breaking it down the cost of FVP over three years could equate to the cost of 2 or 3 New Generation Flash Based Arrays or legacy SAN systems. That’s a hard pill to swallow and it comes down the crux of why I think FVP was priced too expensively for it to gain the market penetration that might have meant this sale to Nutanix could have avoided.

Maybe the exit was always planned and I am coming to understand more and more that startups in todays technology world are governed by the investors who want a return on their risk…VCs don’t pander to what’s cool or what’s great tech…they want their investment to be realized. In this case it looks as though PernixData missed a golden opportunity to get the pricing right, not sell FVP as traditional hardware and penetrate the market more. I know lots of people who would have looked at and implemented FVP if the price was right.

Again, I might be over simplifying the economics of it all but proof is now possibly in the pudding. I’ve blogged about FVP Freedom which was a great initiative to get FVP into as many labs and peoples platforms as possible. Rather than give it away for free maybe the focus should have been on getting the pricing right for FVP. Would have gone a long way to many today not feeling bummed, disappointed and annoyed that a company with such great people and technology may be swallowed up. I am guessing that in today’s tech sector with so many vendors playing in the same space we should come to expect more situations like this…doesn’t make it easier to swallow.

End of the day I feel that the pricing model forced people to see FVP as a purely band-aid only solution that was used only in times of desperation…it didn’t deserve that reputation, but it was widely known as truth both inside and outside of PernixData. Be interesting to see how this pans out after/when/if the deal is finalized. To quote one person …FVP users seem to be the big losers if this goes through.

Investors will always try to get their money!

Released: PernixData FVP 3.5 and Architect 1.1

PernixData has released version 3.5 of their FVP acceleration platform as well as version 1.1 of their Architect storage intelligence platform. Compared to previous releases the list of new features is relatively small however there are a few new features that are worth the upgrade from previous versions.

Architect 1.1:

Probably the biggest addition in this release is the introduction of a standalone PernixData Management Server Appliance that can be used an an alternative to the installed management service that previously only run on Windows. This had been a common request by PernixData clients so its good to see that they have responded and released the VSA. The Management Appliance contains a built-in PostgresDB though it does have the option to connect to an external MSSQL in order to preserve a current config.

Along with the new VSA there have been some improvements to the PDF Reporting and that reporting will now ignore blacklisted VMs that are ignored in the working set calculations. There are also some improvements to the clickability of the VM Performance graphs that take through through to VM performance data when clicked.

FVP 3.5:

For FVP the biggest addition is support for both physical and virtual RDMs though there are limits in that all hosts must be on the latest version of FVP, MSCS and Oracle Clusters are not supported and there are conditions around the use or third part backup applications. Along with that there is increased search functionality for large environments and some better metrics for read and write operations.

Apart from the new features there are a number of general bug fixes and platform improvements as contained in the Release Notes. FVP is an amazingly simple and elegant solution for those wanting to get more out of their existing backend storage or looking to add new levels of performance via SSD or PCI Based Flash…FVP 3.5 and Architect 101 software, along with documentation, is available on the PernixData support portal at  

Quick Fix: Python Error When Removing PernixData FVP Host Extensions

Came across an issue this morning trying to remove old 2.0 PernixData FVP Host Extensions from an ESXi 5.5 Update 3a Host. When running the uninstall script I was getting the error shown below.

There is an old known issue with the version of Python that gets installed with the latest updates of ESXi 5.x and older versions of the FVP Host extensions.

FVP compatibility with ESXi 5.5, 5.1, and 5.0, and Python 2.7
Date announced: March 31, 2015
Upgrade issues exist with various ESXi 5.5, 5.1, and 5.0 releases that upgrade to Python 2.7, which is not compatible with FVP or lower. Upgrading to any ESXi 5.5, 5.1, or 5.0 patch that upgrades the Python version to 2.7 (or later) requires an FVP upgrade to or later. For additional information, please reference KB 1230.

That internal PernixData KB isn’t working at the moment, but I worked out a series of steps to get to a point where the FVP Extensions can be removed. If you have FVP installed with these versions:

  • FVP 1.5
  • FVP
  • FVP
  • FVP

You can apply the same steps to resolve the issue.

  1. Check the Status of the FVP Service
  2. Install over the top of the currently installed version at least
  3. Reboot the Host
  4. Run the Uninstall Script

The Shell Dump is below of the steps above.


First Look: PernixData Architect (Beta)

PernixData had a number of announcements at Virtualization Field Day back in June including details around their v3 release of FVP as well as the announcement of their Freedom Edition. The other announcement was around Architect which is PernixData’s play at deep storage Analytics and Intelligence. With the advantage of being plugged right into the ESXi kernel the FVP Architect Host Extensions allows the software to represent (at a deeper level) statistics around VM performance, IO Profile and working set.

I’ve installed the beta into my FVP Freedom Parent lab to get a feel for what Architect can do and what level of detail it could provide. As mentioned I choose to upgrade my existing 3.0 Manager to the 3.1 binaries you get access to as part of the beta and I have it running side by side with my Freedom Edition.

You can see above that there is a slightly different installer with the FVP Architect Logo and the upgrade process was no different to upgrading between minor or major releases of FVP. After the upgrade you need to go to the PernixData Management Console and activate the Architect license which is set for a 60 day trial.

One thing to mention before moving on is that the Architect platform started to populate data even with the FVP 3.0 Host Extensions installed…meaning that I didn’t have to upgrade to the 3.1 beta bits…not sure what I’m missing out on short term buy not going up to 3.1 but it appears to be doing it’s job ok.

Working through the various Dashboards of Architect you start to get a feel as to what PernixData are trying to achieve with this software…there are the usual performance metrics we are used to seeing (Throughput, Latency, IOPS and Acceleration Rates) but right from the start you see additional bits of information that up until now, have been hard to capture and represent…IO Block Size breakdown distribution and percentage, working sets of VMs and Hosts and then (after an 4-8 hour wait) recommendations are given against VMs as to weather they are best suited for read or write back acceleration.

A few things that I was surprised with from looking through the data was that my NestedESXi Labs where extremely write intensive and the average block size we a tiny 4-8K. Overall I am pretty impressed with Architect and for someone that loves analytics and metrics I find myself being engrossed by all the data that’s being presented and intrigued by what I am finding out about this particular platform.

The beta program for their Architect software is open and you can register below for the 3.1 Beta of PernixData Manager and the FVP Host Extensions.

Below is an overview video taken from the PernixData site:

PernixData FVP Freedom: Setup and First Look

At the Virtualization Field Day back in June PernixData Announced they would be releasing a free version of their FVP software…FVP Freedom. Overnight Freedom was released and as I have been eager to put it to use in my Lab for a while now I’ve gone through the install steps and put together a guide/first look.

As a quick reminder FVP Freedom gives you the following:

  • Acceleration for any virtualized application
  • Up to 128 GB of read acceleration (Using only RAM as Cache – DFTM)
  • Community support

Required Bits for Installation:

  • FVP Management Server (v3.0)
    • 4 vCPU and 8GB vRAM
    • Windows 2008/R2 2012/R2
    • MSSQL 2008/12/14
    • .NET 3.5 SP1 or Later
  • FVP Host Extension for vSphere 5.1, 5.5 or 6.0 (v3.0.0.0)

FVP Management Server and Console:

For my NestedESXi Labs I run five hosts that are a mix of ESXi 5.5 and 6.0 Clusters managed by the 6.0 vCenter Server Appliance. The FVP Management Server is hooked up to the Active Directory for authentication and I am running MSSQL locally on the server to keep the required services local.

Once you have the pre-requisites installed launch the FVP Manager Installer ( I choose Custom Setup )  and work through the install options including configuring the vCenter details and authentication, database and FVP Manager IP and Port settings.

Note: There is still this password issue when special characters are used.

At this point, before installing the Manager you are told that the FVP Freedom Edition is being installed and you can activate using the license page.

If, like me you have installed MSSQL on the same server you may get the following warning.

My suggestion is to cap MSSQL to allow for the 4GB free memory the FVP Manager is asking for…the Management VM has 12GB of vRAM and I’ve limited MSSQL to use 6GB max. Once the installation is complete you can now access the new standalone FVP Management Console. If you choose the default options you can browse to the Manager Console via:


Once logged in using the credentials supplied during the setup process you will see a familiar layout to what was previously in the VI Client and will show up in the Web Client.


License Install:

To install your FVP license (you should have received the license in a seperate email from PernixData after registering for FVP Freedom) click on the Licensing Tab and select the vCenter Server from the list and Click on the Enter FVP License Key.

At this point the Online Activation should take place and you should see the FVP License Type change to Freedom Edition. Clicking on the License details in the lower pane will show you just want PernixData are giving away with this edition.

FVP Host Extension and Host Resource Configuration:

Installing the Host Extensions is a straight forward process and I won’t go into too much detail as there are other blog posts around that cover the subject. In a nutshell you can use Update Manager to push the bundle out or use the following esxcli command on the hosts directly.

esxcli software vib install-d /vmfs/volumes/ds01/

For the rest of the FVP Cluster Configuration I’ve created a quick video:

vCenter Web Client:

With the new Stand Alone Management Console the vCenter Web Client is used as a pointer. As shown below the PernixData Logo now shows up in the Home Page and clicks through to a summary screen where you see the FVP Cluster details and get some basic insights into how the cluster is performing.

Enjoy the Freedom!

Released: PernixData FVP 3.0 – New Console + vSphere 6.0 Support

Last week at VMworld I was lucky enough to spend some quality time with the PernixData team and got some great insights into their future products including FVP, Architect and FVP Freedom. Today PernixData has released into GA FVP 3.0 which includes some significant new features and supportability for all those who are on vSphere 6.0. There are also some performance enhancements, support for offsite metadata analytics of if you choose to allow PernixPlus as well as an improved license activation experience through the new HTML5 Portal.

Support for vSphere 6.0

FVP 3.0 introduces support for vSphere 6.0.  Maintain your world class performance as you transition over to vSphere 6.0.  FVP 3.0 also supports mixed environments of vCenter 6.0 with hosts running ESXi 5.1 or newer as you phase in vSphere 6.0 across your data center.  With the support of vSphere 6.0, FVP 3.0 will no longer be supporting vSphere 5.0. 

New HTML5 based User Interface

FVP 3.0 offers a completely new user experience.  Administer FVP using a modern, fast, and flexible interface with a web browser while moving past the inherent limitations of the vSphere Web Client.  This new interface can be used on its own, or easily launched from the vSphere Web Client.  FVP is known for its elegant presentation of performance metrics, and version 3.0 improves on this experience even more with enhancements to the performance data presented.

FVP is an amazingly simple and elegant solution for those wanting to get more out of their existing backend storage or looking to add new levels of performance via SSD or PCI Based Flash…FVP 3.0 software, along with documentation, will be available on the PernixData support portal at  

As a sidenote, with this release PernixData have stopped official support for FVP 1.0.

PernixData – Giving FVP Away! #VFD5

At Virtualization Field Day 2015, PernixData CTO Satyam Vaghani presented to the VFD5 delegates on some of the new features being released by PernixData. Personally speaking FVP is already a great product and at times I wonder what more can be done to make it better. However from what I have heard and now seen at VFD5…PernixData are not going to rest on the current success of FVP.

In what is becoming less and less of a surprise these days with disruptive Tech startups PernixData are releasing a free version of FVP called “FVP Freedom” This will be a free, community based version of FVP.

The free edition will come with the following limitations:

  • One cluster only
  • DFTM Only (no SSD or PCIe acceleration)
  • Read Acceleration Only – 128GB Per Cluster write-through
  • Community support only

The fact you can only accelerate VM read workloads on RAM is a little bit restrictive (and resource expensive) but being able to use the DFTM-Z feature is seriously impressive and a very smart move by PernixData who openly state that they way FVP in every ESXi hosts on the planet! Grand plans forsure, but by releasing this free tier it allows enthusiasts to consume the product at it’s most capable.

For a further read up on the reset of the PernixData Announcements at #VFD5, head over to Duncan Epping‘s blog post here or have a read of James Green’s post here.

For those that are interested, you can pre-register for FVP Freedom here:

Additional Links:

Fully functional 30 day trial of FVP 2.5 here:

At Zettagrid, we have already taken advantage of what PernixData has to offer and have integrated the FVP solution into parts of our IaaS platform. Zettagrid Case Study:

Released: PernixData FVP 2.5

PernixData have GA’ed FVP 2.5 and it’s got a number of enhancements over the previous 2.0 release. For those running ESXi an haven’t heard about FVP here is a quick summary:

PernixData virtualizes server-side flash and server RAM across all hypervisor nodes in a compute cluster and hooks the high-speed server-side resources into existing VM I/O paths to transparently reduce the IOPS burden on an existing storage system. Customers leverage PernixData’s FVP with their existing primary storage deployments and manage them within the context of their familiar hypervisor management tools

In a nutshell it’s awesome and if Veeam hadn’t recently patented a certain phrase, I would associate that particular phrase with FVP.

New features in the FVP 2.5:

  • Distributed Fault Tolerant Memory-Z which compresses data stored in RAM.
  • Intelligent I/O profiling, which provides a way for administrators to temporarily suspend and resume acceleration on virtual machines, and without deleting the flash footprint for those VMs.
  • Role-based access control
  • Network acceleration for NFS datastores.

For me the standout being Memory compression which opens the door for more efficient use of FVP to use precious/expensive RAM resources as acceleration cache. Check out Frank Denneman’s blog here where he goes through the feature in detail. Some of the numbers accosiated with FVP RAM acceleration are just silly, so the ability to compress and be smart with the data used in memory makes it more attractive to add a little more RAM to those Host Build sheets.

The I/O Profiling is manual at the moment, but is a good start for those who know VM workloads that could potentially inject “dirty” blocks into the FVP cache during backup or Virus Scanning operations…I’d like to see this become even more intelligent at become self aware of data that’s dirty…this would be a big help for Service Providers.

In terms of Bug Fixes, there is a fix for those using Using Veeam to restore a VM to a different datastore on the same host which had resulted in the VM remaining in a stalled state.

Thanks for PernixData who made me a PernixPro last week…I love the technology and it’s certainly done wonders within our ESXi platforms to help solve issue with higher than wanted storage latency. If you have spare SSD or want to check out the RAM cache features of FVP, head to the site and download a full 30 day trial…easy to install and low impact in terms of configuring FVP against datastores or VMs.