back to article Partners? With them? No way, says HPE of Nutanix

HPE has taken the very unusual step of pointing out it is not partners with Nutanix. The hyperconverged upstart last week announced that it is now possible to run its software stack on HPE servers. That announcement took HPE rather by surprise, because it turns out Nutanix hadn't asked in advance. HPE's clearly a bit miffed …

  1. Anonymous Coward
    Anonymous Coward

    No UCS support. Nutanix got kicked out of Cisco's Partner Support Program for a similar stunt they pulled with UCS support. Cisco hates their guts. Come to think of it, so does everyone else. Before you think it's because everyone is afraid of them, it's because Nutanix is a shady player in the industry. Half-baked products that are touted as the best thing since sliced bread, driven by an aggressive sales organization that bamboozles gullible customers, constantly skating along the edge of legality or ethics. They are the worst example of the bullshit until you cash out silicon valley culture.

    1. wjthompson

      Get burned?

      Curious, has Nutanix burned you too?

      Yes, they're pulling some shady stunts with these compatibility/certification shows (customers would expect this to be a two-way trust between software and hardware provider), but this is fundamentally because customers are demanding it, and the OEMs are trying to prop up their own hyperconverged platforms.

      Shit though, this is software defined. Certifying on known configurations is step one before allowing it to be run on the broad spectrum of hardware in the world which I'm sure is Nutanix's next goal - and they'll be leaps and bounds further than anyone else when that time comes.

      1. Anonymous Coward
        Anonymous Coward

        Re: Get burned?

        You mean, more than VMware (vSAN) or Microsoft (Spaces... stop laughing at the back).

        Leaps and bounds nearer the other two pure software ones, then.

      2. Anonymous Coward
        Anonymous Coward

        Re: Get burned?

        They have bullshitted us constantly. It got so bad that we had to tell them to never call us again. We also severed our relationship with a partner who tried to push them aggressively. Are the partners getting bribes or what's the deal here? Right now our policy is, a partner who mentions Nutanix goes straight on the no-buy list.

        We are a medium sized shop, Cisco networking, HPE (mostly older) and Cisco compute, 3Par (also older) and Pure storage. We have some VSAN in production, just deployed some HyperFlex for a smaller environment, but are generally happy with our Pure array. We run Vmware, did vRealize but switched to UCSD, mostly for the self-service catalog and chargeback. Our developers run Ansible and started playing with Docker. We use CommVault for backup, HPE Apollo as targets.

        Nutanix came in early and tried to sell us their Acropolis appliance. We told them to take a hike as we were in the middle of a vRealize project (which ultimately failed due to a weird bare metal app). They just kept coming up with one easily disprovable thing after the other. Look, I don't want to claim that we are some super gurus but we run a pretty tight shop here. We got the impression that they play fast and loose with facts and that they are taking advantage of less experienced folks. The more we looked into their tech and their claims, the less we liked it.

        They kept coming back and every time they had a new feature that upon closer inspection turned out to be utter shyte. Cloud mobility? Poor man's version of Veeam's cloud backup. We challenged the sales idiot and his idiot systems engineer to explain to us how they stretch Layer 2 into AWS. When they realized that we were unwilling to switch away from UCS they claimed their "enterprise support" for Cisco. Come to think of it, they constantly use terms like "enterprise" and such when in reality they are all smoke and mirrors.

        To sum it up, the guys we worked with were a bunch of sleazy con artists and nothing that I've seen from that company made me change my mind. Every pseudo-independent blog post I read turned out to be from an employee and is always just a little bit too much. My bullshit radar is just going off constantly with these guys.

        1. Anonymous Coward
          Anonymous Coward

          Re: Get burned?

          Yours is a "tight shop"?

          More like tight-wad.

          More like tight necktie. If you loosen it you might notice the blood returning to your head.

          Let's get this straight, your so-called tight shop runs HPE and Cisco servers, 3Par and Pure storage, and VSAN and Hyperflex HCI. Right?

          What damn fool is running your shop?!?

          I can see why the Nutanix team was after you because you are the epitome of inefficiency and the reason HCI is taking off like it is.

          1. Anonymous Coward
            Anonymous Coward

            Re: Get burned?

            I see your social skills are top notch.

            We used to be a large HPE and 3Par shop and still have a few legacy systems running on the remainder of C7000. That's our old ERP, a few Oracle, and various partner/supplier interfacing systems that haven't been fully migrated into SAP HANA and Oracle 11g. As soon as that finishes we are decommissioning the blades and the 3Par.

            We have 14 UCS chassis with 106 blades in prod, most of it virtualized on ESX with a few bare metal Linux boxes. All our CentOS for the DevOps guys are rackmounts, so are our SAP HANA servers and the Oracle RACs. Everything on two Pure m50 and one m20. That's without the 55 node Big Data cluster on UCS rackmounts that's being rolled out now. The data geeks say it's already obsolete because they are evaluating Nvdia for Tableau which is allegedly 1000x faster than anything on x86. Not my area of expertise so I don't know.

            The VSAN cluster is from the time when we hadn't deployed UCS and Pure fully and the developers were still coding on VMs. They constantly requested new LUNs and we couldn't get vRealize to work properly with the 3Par so workflows and self-service were a no-go. Ops management felt uneasy about giving the developers admin-access to the array. So we just installed VSAN on four rackmounts and called it a day. They got their little DEV environment and we didn't have to bother with them.

            The HyperFlex is nice and we got a sweet deal on it but to be honest, we already have Pure arrays and don't see much use for the HX. Right now we run it as a admin pod but it hopelessly oversized for that, we might just ship it out to DR and test Oracle on it.

            I don't know wtf you think a tight run environment is but we have ~8,000 VMs and ~50,000 containers under management. We have self-service, workflows, SDN, DevOps and CICDCD (for the microservices), layer 2 extension into AWS and bursting into EC2.

            Maybe you want to share what you do all day before you start criticizing folks who actually do this for a living.

  2. Anonymous Coward
    Anonymous Coward

    vSAN and Microsoft as pure software? They have very strict HCLs, how is this ANY different? I suppose VMware leaves it up to all the hardware manufacturers to figure out what kind of hardware they want to tell customers is vSAN Ready. Ask anyone who has gone through a major code release of vSAN they will tell you when those HCLs change they get burned. Sounds like Nutanix is taking the time to actually test the hardware and recommend configurations to their customers. Kind of refreshing, really.

    1. Anonymous Coward
      Anonymous Coward

      vSAN HCL's change?

      vSAN HCL's don't change with each major release. There is this myth that stuff gets dropped with every release and outside of the original release dropping some HBA's I'm not aware of anything actually being purged from the HCL.

      1. 2 years ago, From 5.5 to 6.0 SATA magnetic drives were not kept going forward (too slow, and some drive vendors plaid fast and loose.

      2. 2 years ago, some HBA's were deprecated for being too slow.

      3. With the 6.2 launch Dell was 2 months late to being certified as they worked to fix some issues that were identified in testing.

      4. with the 6.5 launch HPE was 3 months late to being certified as they worked to fix some issues that were identified in testing.

      5. Other than that anything not kept on the HCL going forward tends to be obscure hardware, or the server itself not being kept on the vSphere HCL (HPE Gen7's I don't think were certified for 6.5 in general, etc).

      Note, VMware requires a joint statement because otherwise they would have zero leverage to get vendors to fix driver/firmware issues. This is something I don't know how Nutanix intends to fix given their status as a non-supported OS on these platforms.

  3. MikeR

    Call me cynical but

    The Nutanix HP 'announcement' seems to have served its purpose... it's got folks talking about Nutanix. Twitter is alight with many of my ex HPE colloquies rebuffing the statement - and as such talking about Nutanix - well played Nutanix, well played...

    P.S. I don't work for Nutanix.

    1. Anonymous Coward
      Anonymous Coward

      Re: Call me cynical but

      That's a good point. I hadn't even thought of that.

      Well, Nutanix runs on Cisco even though they aren't an OEM. All you have to have is the type of h/w Nutanix tells you to have. So, h/w failure, call Cisco. Software failure, call Nutanix.

      Now, Nutanix has the same thing for HPE.

      I see a pattern developing here.

  4. Anonymous Coward
    Anonymous Coward

    Toys and prams

    If you sell your own HCI software, you're not going to like someone else installing their software on your hardware. Cisco didn't like it either.

    Notice how Dell and Lenovo dont give a dam, because it actually gives them a HCI offering to go to their customers with?

    Personally i love the fact that Nutanix have gone along and verified (for themselves) that their software will run on other vendors kit and are prepared to support customers making that choice. At the end of the day is not special, it's just x86 servers running a variant of KVM and I'm fairly sure HPE have a partnership with RedHat that already certifies that KVM runs on ProLiant servers.

    1. Gerhard Mack

      Actually, Dell has their own HCI offering.

      1. MikeR

        'Dell has their own HCI offering'... actually we have a few and they are all rather good, helps us to help our customers put their technology on the right platform, which could also be Converged or 'build your own'

        *Iwork4Dell

        1. Gerhard Mack

          @MikeR

          "actually we have a few and they are all rather good, helps us to help our customers put their technology on the right platform, which could also be Converged or 'build your own'"

          Yes, I know, your people stopped by my workplace and treated me to roughly 2 hours of presentations on the topic. Your president of Converged systems really likes to talk.

      2. Anonymous Coward
        Anonymous Coward

        Dell's HCI

        True, it's call EMC's vxrail. But, like most of the other products on the market, it's a gen1 or, at best, gen2 product and still considerably inferior to what Nutanix offers.

      3. Anonymous Coward
        Anonymous Coward

        I fairly sure they purchased EMC AFTER the partnership with Nutanix. Also VXRail only works with VMware. The partnership enables Dell to sell to HyperV and KVM/AHV customers too.

        The partnership works for Dell, at the moment.

  5. Anonymous Coward
    Anonymous Coward

    What's shady?

    What's shady?

    These incumbent, old-school, vendors claim their servers are so great and work so well. But, wait, you can't put competing software on our servers! No, that won't work at all. Our servers will only work with OUR software.

    It's all crap. The end is near for proprietary h/w vendors.

    1. Anonymous Coward
      Anonymous Coward

      Re: What's shady?

      It's not shady.

      Having worked with the OS certification teams at the major OEM's I can tell you it's expensive. It requires testing teams, who regression test every configuration and every OS build. It requires teams to troubleshoot issues between the OS and hardware firmware (Sometimes this requires paying 3rd parties like the OEM who manufactured the device to fix issues, or check out code). Once new firmware fixes are shipped they have to be regression tested and certified also. It can easily be a 9 figure commitment once you include PM and other overhead, and doing this for a platform who promises to not share much revenue with you and undermine your other products would be stupidity of the highest order.

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like