Nutanix

<SmellyHippy>
Because Information wants to be free man.
</SmellyHippy>

I think its unreasonable to only allow one side of a story to be told. I think its unreasonable to want control of 3rd party conversations regarding your product.

By all means, Rebutt to your hearts content, I'm sure $Vendor has more reach than $Blogger.

This thread was started on the 4th of October 2013, on the 5th of October 2013, there was a Nutanix rep posting (who had an account here for a long time, but never posted). It's obvious companies already scour the web for references so they can control the conversation. So yes, I do believe these companies have time (and budget) for this.

They don't need to get someone to publish a retraction, http://blogs.vmware.com/ and http://www.nutanix.com/blog/ exist, and could be used for debunking spurious claims.

With such strict controls in place around conversations, it becomes increasingly difficult to find information regarding anything 'negative'.

When you need to look at any benchmarking of VMware and Nutanix through glasses that have a 'This benchmark was approved by VMware or Nutanix' filter. Making decisions based on those benchmarks is far less than Ideal.

If the product and/or service is good enough to stand on its own 2 feet, then let it. Until then, I'll regard benchmarks from those products with the same skepticism I regard Media releases from Inside North Korea (is Best Korea). .

 

Every other business seems to survive, why do they thing they're different?

Actually, in the case of VMWare especially they have a marketing team doing just that. They manage all social perception of the product. Nutanix I'm not sure.

It happens all the time. Nobody sees it as a credibility hit at all. Reviews are edited and you'll see them post the update transparently. It's really not hard to find examples of this. Here's the first example I can find: http://arstechnica.com/gadgets/2015...e-support-play-services-data-backup-and-more/

The journalist thanked and referenced the rebuttal post about it.

Have you ever received a "cease and desist" letter? I have. You need to remove all references to the subject / company or you'll be considered to be in breach.

Now my question to you, how would you feel if the federal government legislated that you must run all NBN reviews and articles through their team before being allowed to publish them? Please, be honest here.

 

That's not actually true, in fact Nutanix would be more then happy to pull the vetting process from the Eula if VMWare did so as well but they aren't that keen.

Yes the VMWare / Nutanix relationship gets ugly at times but honestly as a customer of both I see the horse shit from people like Chuck Hollis as just that horse shit it's open vapour without substance meant to either bait a response from some Nutanix employee's or spread FUD.

VMWare has been very aggressive toward Nutanix for over a year now, I would actually rather they just got on with things. Seems they want to play the man not the ball. They also have a habit of threatening people to remove content they don't like, ie the first EVO:Rail reviews. EVO was launched last August and yet there a serious lack of content.

I am still in Miami wrapping up from the .NEXT Conference, I was here as a customer and as an NTC to launch Community Edition. On the CE subject you can now download CE at http://www.nutanix.com/products/community-edition/register/

I have a number of invite codes I can hand out to Australia's if you should want one.

Btw all the one click stuff we saw this week will also go the other way, ie kvm or hyper-v to esx. Customers like myself who are a VMWare heavy shop are not being left behind, we are being given more options and that is always a good thing.

 

Why are Nutanix beholden to what VMWare do?

Microsoft don't care if I (or anyone else) publish benchmarks of My Hyper-V cluster running on Raspberry Pi's, using IP Over Avian transport, back to my JBOD of floppies, Neither Do Citrix if I do the same with XenServer.

The whole Chuck Hollis/Josh Odger's pissing match reads like a script from Home and Away. If you know that someone is baiting you... then common sense says Don't Bite.

I would never had read anything that Chuck Hollis wrote, if Josh Odgers didn't write (and heavily link) counters to them, so in a way, it's kind of self defeating.

No sane person reads a Vendors blog and thinks "This is a fair and impartial comparison between Vendor and competitor"

Its good that VMWare customers aren't going to be left behind, but from an outsider point of view, I'd say one can be forgiven for thinking that from a VMWare shop, Nutanix is probably not a good direction to be headed in, and From a Nutanix Shop... ESX probably wouldn't be the first choice of Hypervisor for you.

Why would I wan't my infrastructure split amongst two vendors who can't talk to each other without it degrading into a slanging match?


I was going to check out that link for CE, as it seems like a good way to go (especially if Nutanix permit people to publish benchmarks), but then I got a full page splash screen and thought my broswer had been hijacked .

"Hey, We should have a full page splash screen when anyone visits our site, people love that shit"

For a company that does some pretty slick things in the Tech marketing space... whoever greenlit that one must have been having a bad day.

 

Yeah not sure what they were thinking, but that splash is new - it wasn't there when I registered a week or so ago for CE, and it didn't show more than once for me.

Still a misstep though, I think.

I'm interested in seeing whether I can run up CE on a couple of little Atoms to start with - I have a pair of old Microservers. Sort of torn between wanting to try "the MS way" with Server vNext or the Nutanix way with CE, for the home system/labs. And I don't have the ready for 8 new hosts!

 

Just following up on the free Community Edition. Send me a PM if you want an invite.

You can run CE nested on ESXi, Workstation or Fusion too to play around with it (if you don't want to use hardware).

 

How picky is CE with hardware? Will older ibm/hp servers be ok? Will it cry if you don't have huge disks?

 

Minimum System Requirements

CPU : Intel CPUs, 4-cores minimum, with VT-x support
Memory: 16GB RAM minimum
Storage Subsystem: LSI controller (HBA, JBOD or RAID 0 mode) or AHCI
Hot Tier (SSD):Single 200GB minimum SSD
Cold Tier (HDD): Single additional 500GB minimum device
Networking: Intel NICs recommended
Boot Device: 8GB minimum, USB3.0 media recommended

Once you get access to CE, jump on the community forums at my.nutanix.com - people post their setups there.

diomedes on ocau forums would know more than most on what's the latest update re h/w support

 

On the 'No negative reviews allowed front'

What motivation gets assigned to this.

http://www.storagereview.com/why_we_don_t_have_a_nutanix_nx8150_review

 

Interesting to say the least, anyone reading that article would think twice about purchasing Nutanix.

 

For me, the interesting part is the public airing of it, Normally this sort of talk is all secret squirrel stuff, and then end result is an article that all parties are 'happy' with.

For a company with the technical nous of Nutanix to let it get this far is massive. The only real way back from this, is to ship StorageReview some kit, with the brief of 'Do whatever you want with it'. Trying to lead the HCI charge by providing the same staged, synthetic benchmarks as the other players in the field is self defeating.

Change the game, Let reviewers do whatever they want with/to your stuff, and then you can point to all the other companies and say 'Our independent benchmarks actually mean something, yours are staged'... In my mind, the only reason not to do this, is if you aren't confident in your product.

I'd trust the company that has worse, but honest numbers, of the one that tries to blow smoke up my ass with high, but utterly irrelevant ones, and I'd say most people in my situation are the same. The tech community on the whole, is pretty skeptical, and doing shit like this just playing into the whole shitstorm.

 

Not sure that it'd put that much doubt in someone's mind, but it's not a good look at all. Nutanix have (what appears to be) one of the best communities and customer driven success stories so seeing them act in the way they did isn't good.

They didn't do anything overly wrong (vendor specific benchmarks are always biased so it's a good thing to question them), it's simply the manner in which they did it. It's exactly why Brian went public about how it was handled and also why Nutanix themselves have issued an apology.

We (the tech community) need independent tests. Reading whitepapers from each vendor stating how good they are is a complete waste of time and the top tier vendors are very bad at it. Two of the recent comparisons I saw was comparing Vendor A's gear to 3 generation old Vendor B gear. As you can imagine, there was a clear "winner" but a Luke212 whitebox solution would have achieved the same thing.

In this case, people are going to choose between running a pure VMWare solution or a hybrid VMWare / Nutanix platform. Those looking at the systems would greatly benefit from comparisons and if either vendor is upset with the way in which the tests were run (ie VMmark), then write a follow-up article. No doubt it'll rank highly SEO wise from the vendor itself, so users will find it. I'm sure the original author of the results would post the changes / update their article too, especially if they're independent.

 

I have to admit I've gone off the idea of seriously evaluating Nutanix CE for now, for two reasons

1. The fact that CE requires you only use KVM - all my VMs are Hyper-V and I don't plan on converting a stack of VMs for a comparison test;
2. Very, very recently this Storage Review kerfuffle.

I'm a bit disappointed really - item one isn't necessary (by all means point the user to the free Hyper-V and ESXi downloads rather than including them). One of their big selling points is that you can run any supported hypervisor - unless it is CE, in which case they seem to desperately want you to use the Linux hypervisor (is it so they can para-virtualize and the performance looks better?)

Item 2 is more worrying. Any vendor should have right of reply - and Nutanix have that. But they should not have right of veto in this fashion - it stinks of cherry picking and favouritism.

Edit: If you read the reply from Lukas in the SR forum, I don't even think he's being too unreasonable about wanting to use the same hardware - but that being the case, ante-up the same systems as the ESX guys are using, or accept that you sell a different platform and request that readers are made aware that "VSAN chose to use 1.2TB 10K drives and Nutanix 1.0TB 7K drives, which means for the same number of spindles you might expect the VSAN solution to be able to deliver up to 35% more IOPS". They also have the right of reply on the basis of performance differences due to hardware. But here's the rub - they don't sell NX's with SAS 10K drives that I can see. So tough luck - your SATA system gets compared to the 10K SAS VSAN because that's what the customers will end up seeing/purchasing/comparing. When they sell SAS 10K NX's they can retest.

 

The point of CE is to showcase Acropolis - which is Nutanix's free KVM based hypervisor that they will now concentrate on developing to make the platform not only a VSAN but a VMware killer. Not saying this is anytime soon, but there are a number of large users that have already adopted this and as features are added in the coming months the case will become more and more compelling.

I can see the appeal of the saving of huge amounts of vTax by moving away from ESXi, but of course you have to take everything into account and losing Zerto support for my DR may end up costing me more than I save, so it could be I end up going to Hyper-V after all.

CE is doing its job though by scattering experience of the product far and wide. A year ago people would have said "who?" when you mentioned Nutanix, now a lot know of somebody who's implemented it if they haven't themselves.

What lies in the year ahead?

 

It's not scattering the experience enough though, purely because it is limited to KVM.

If I'm Vmware or hyper-v shop, and was investigating HCI, then Nutanix CE won't even gt a look in, because it wants me to change Hypervisors. If CE was a community edition of their hypervisor agnostic solution, I'd probably already be playing with it.

What lies ahead, no doubt more shit slinging between Josh Odgers and Chuck Hollis for us all to laugh at, and still no real independent benchmarks to speak of.

 

If you're that tied to a vendor's hypervisor, then you're probably not the type to investigate emerging technologies anyway. They've aimed it directly at tech enthusiasts who want to see how it works, which means that the underlying hypervisor isn't as important. Nutanix have V2V tools as well, although I don't know if they're available in the CE edition.

In that case, you wouldn't be comparing nor evaluating the community edition, just the same as you wouldn't evaluate VMWare Fusion / Workstation for the vShpere type deployment. They have trial licenses for the software only offerings, so it's not like the CE variant is the only option.

 

That's a fair comment.

If i get a spare couple of hours on the weekend, I can spin up a Hyper-V or VMware on a whim, and play around with it; What's the best way of doing similar with Nutanix as the platform?

Microsoft will give me 180 day evaluation versions of almost all their products with pretty much just an E-mail address, VMWare are similar (although 60 days).

Nutanix seem to want to Solarwinds me to death with Sales representatives and account managers, before I can even look at the real thing. That probably works with their customer focused business model, but it certainly makes dipping a toe in the water, a daunting proposition.

 

I've yet to even try it fully, but there's this blog about running it in a nested VM arrangement: http://next.nutanix.com/t5/Community-Edition-Blog/bg-p/Blog

Remembering that you're not just looking at a hypervisor, so it's not exactly an "on a whim" type product to even play around with. I still a hate the "hyperconverged" term as much as cloud, but you're looking at many pieces of the whole infrastructure puzzle in one box / cluster.

 

This sort of loops around to the "problem" of it being KVM only. You can't eval the management UI properly without creating and managing datastores and VMs (and all that that entails) - and then actually RUNNING the VMs. You can't compare how a "representative" system goes and really play with pulling disks and SSDs and moving VMs around, etc.

You most certainly can't break it and fix it (there's just not enough "hardware").

I was actually considering trying to get my hands on a cheap Dell C6100/C6220 or similar (3-4 blades in a system with 3-6 disks each) to play with it properly, but the combination of "buy hardware" with the thought of "V2V all/some VMs or rebuild a stack of them" is putting me off. If I could use the existing VMs, or at least migrate them on/off - it'd be much more attractive.

 

You know the best way to make blogs nobody reads... Require registration .

There is no mention of the Storage Reports article on /r/Nutanix... despite it being very relevant to the platform.

I'm was a big fan of how Nutanix used to engage the community, back when they were 'the little guy'. But the pissing match with VMWare, and the desire to control the conversation regarding their product/s haven't really done them any favors in my book.

I've got hardware, and while parts of it may not be on the HCL, VMWare, and Microsoft will still let me install and run fully functional trials of their software, Which they will provide trial keys for after filling out a basic online form. This is where the on-a-whim factor comes in. If I'm looking to do the same over a weekend with Nutanix, I seem to need to pre-arrange all of this and have things vetted by an account manager or sales team. If I'm mistaken, and I can just fill out a form, and download a trial version of Nutanix, send me the link .

I much prefer the full trial version over a feature limited 'community edition' as a way of exposing tech enthusiasts to technologies, I'm not sure how many other tech enthusiasts share the opinion.