HPE VM Essentials

[JaySt]

Thats indeed something they also mentioned, about the connection between the teams. So all good i hope.
Lets go!

[Gostev]

FYI today Veeam has officially announced the intention to develop a native integration with HPE VME
https://www.veeam.com/company/press-rel ... overy.html
https://www.hpe.com/us/en/newsroom/pres ... overy.html

[JaySt]

Good news!

[Didi7]

@HPEStorageGuy

I really must admit that HPE VM Essentials looks like an interesting virtualization target platform for our future onprem infrastructure environment. There is hope to finally move away from this Broadcom VMware debacle. Nutanix did not convince us as well, we used it several years apart from other VMware infrastructure environments and imo its also way too expensive, less mature and more complex than VMware.

Nevertheless a free Community Edition of HPE VM Essentials (lets say 10 workloads) might draw more attention to the product. See what happened to Veeam since they launched their Community Edition, imo it was a very wise decision from Veeam. So spread the word!

No need to mention that it is awesome news that Veeam has decided to fully support this remarkable new hypervisor environment!

Regards,
Didi7

[Gostev]

TIL (from Reddit) that HPE VM Essentials is only supported on HPE hardware. Feels like an artificial limitation to drive more HPE hardware sales, however definitely something to consider when planning its adoption, depending on your existing server and storage fleet.

[JaySt]

correct, currently. But that could change. I think it's already on the roadmap to have it on other hardware as well.
HPE is also (preferably i think) packaging this as a part of disaggregated converged infrastructure, together with compute and storage, managed as a whole from greenlake cloud platform. They call it PCBE (Private Cloud Business Edition).

[CalvinZ]

@Gostev and @JaySt: Several 3rd party servers and storage are currently in test. Obviously, it's far easier for us to test HPE gear but going to 3rd party gear was always on the roadmap and is now in flight. All HPE Private Cloud solutions will include our HVM (the HPE KVM instance in HPE Morpheus Software). That includes HPE Private Cloud Business Edition and HPE Private Cloud. Also coming is HPE SimpliVity with VM Essentials. VMware will still be an option for customers that want to use ESXi as their hypervisor.

@Didi7: I'm working on a HPE Morpheus Enterprise Community Edition (branding won't like me but I'll call this MECE). Not sure what it will be but it was initially defined as 4 sockets. To allow home-labs with external block storage and dual-socket CPUs, I'm trying to get approval to have 6 socket MECE licenses. Not sure of the timing but hoping it will be in the next 4-6 weeks. Would love to hear back from home-labbers who want a Community Edition, how many servers, how many sockets, and do you have external block storage. Note the 3-server requirement is because KVM uses GFS2 and that requires quorum.

[CalvinZ]

BTW, here's a link to the HPE Morpheus VM Essentials Compatibility Guide: https://support.hpe.com/hpesc/public/do ... 05501en_us. I spy a couple of Dell servers have made the list (and again, more coming).

[Gostev]

@HPEStorageGuy for free offerings I recommend that you consider something like the current Veeam approach: say 2 sockets CE (with no production usage restrictions) and 6 socket NFR (for non-production usage like home labs). That's just something I learnt the hard way: if you make CE too good, the business will want to kill it off anyway, and it will be permanent overhead for you defending the offering in related quarterly calls (this has been soon 10 years of my life across three different management teams). While much larger NFR is never seen as a threat because it cannot be deployed in production and is for labs only.

Also, if it's not too late, do consider Core-based vs. Socket-based licensing right out of gate. Otherwise you're destined to some painful licensing revisions in 5-10 years from now, when a single socket will have gazillion of cores, even your larger customers only needing like 10 sockets total for their entire environment and your business faces dramatic drop in renewal revenue.

[jhoughes]

@HPEStorageGuy, MECE... I like it. For a community edition, I'd love to see 2 sockets, and for an NFR maybe 4 or 6. (Or a vExpert/community champion/, evangelist level of license whether per socket or cores)

I've just started getting it back in the lab, as I hadn't worked with it in a few years. I'm planning to release some unofficial documentation about getting it working with Pure Storage (doing it as a "my hat is off, this is vendor agnostic, just showing everyone their options" type of write-up).

Either way, hope you're doing well dude, and glad to hear you got saved. It was stupid to see you get your gold name badge and still be trying to find a role...

@Gostev, we really appreciate you supporting the fight for CE over the years. I still run the CE within 6 environments that I support, including the non-profit that I help lead, and another that I support. Please know that it's a decade of a good fight, for those of us that work with you, and those that don't (or just not directly anymore).

[CalvinZ]

@Gostev - thanks for the feedback. I'll add your feedback to my list of things as I work through it. We have NFRs for partners that have higher socket count and customer POC/trials but I'm not managing those - just the CE. The intent of the CE is non-commercial home lab use only. I'll raise the idea of a CE for commerical use but my role is focused on community.

The cores vs sockets was a decision made by HPE execs across all HPE Morpheus software to have VERY competitive pricing for paid licenses. For example, the HPE Morpheus VM Essentials costs $600/socket. I don't think I have an option to have the MECE licensed by cores since the baseline is sockets.

@Jhoughes - Yo Joe!!! Nice to hear from you. No comment on the knuckle-heads that thought it was a good time for me to get a gold badge.

I've had a handful of home-labbers tell me that for their homelabs, they want 6 sockets so they can use external storage with 2-socket servers. Not sure I get what you're thinking - 2 sockets on a home lab CE seems pretty limiting (one or two servers max). I'll reach out to better understand. The NFRs that partners or ISVs would use in their "official" labs would be 12-16 sockets and I'm not doing those myself, just focused on a CE.

[jcollins1]

We are officially migrating from VMware to VME. Hoping for some light at the end of the tunnel regarding VEEAM here.

Edit:

Just saw the post! Awesome!

[mark.salomon@acp.de]

Any news on the snap integration / vme native Veeam integration? The market is waiting for this to be released, then we can address our customers with this solution. Without a vaild backup integration (no agent backup is not enough) we are all waiting... Would love to see pressure on this topic!

[Gostev]

No news since last week see above

[CalvinZ]

@mark.salomon@acp.de - As Gostev said, nothing more since what was announced at Discover. But to recap, HPE and Veeam announced that we will have deeper integration between HPE Morpheus VM Essentials and Veeam: https://www.hpe.com/us/en/newsroom/pres ... overy.html. Now the work needs to happen. The HPE Morpheus team has been working on Change Block Tracking API and Veeam will take it to the finish line from there.

[bheungens]

Meanwhile it is september and no further news from the Veeam side...
HPE keeps on releasing updates on VM Essentials making it ready for me to start migrating... Only thing that stops me is Veeam image based backup support...
Don't understand why it takes so long since Proxmox is supported which is using the same KVM hypervisor.
Any thoughts or insides @gostev

[Gostev]

"Takes so long" comparing to what? How much do you think implementing Proxmox support took us?

In general, almost everything these days is "using the same KVM hypervisor" but the hypervisor management stack is always different, so the integration complexity depends almost exclusively on the maturity of integration backup APIs provided by the hypervisor.

Proxmox did not have them at all and thus the integration took quite long, while Scale had a very mature backup API and we were able to integrate with it more than twice as fast. HPE is somewhere in between, closer to Proxmox actually. For example, if you read this topic, VME does not even provide a purpose-built Changed Block Tracking API, which is the absolute essential prerequisite for protecting environments with more than a few VMs.

Because of that limitation, some of our competitors decided to ship a "checkbox" integration without change block tracking and that would be something we could do fairly quickly, however we don't do checkbox integrations just to win RFPs. When we decide to invest into building something, it has to be usable in real-world environments and at scale.

[JaySt]

did HPE already enable the CBT API for Veeam to work with? they're at version 8.0.9 now i think.

[Gostev]

No, they did not. So our current plan is to try and use a low-level API of the underlying KVM directly. Similar to how we approached this with Proxmox.

[JaySt]

ok. that's a bit of a setback i think.
I hope it's doable for you guys.

[Gostev]

As of now it looks promising.

[CalvinZ]

No, they did not. So our current plan is to try and use a low-level API of underlying KVM directly.

I think this might create a misunderstanding - have already heard from someone asking me about it. CBT and a CBT API is and always has been part of the HPE VM (HVM) that of course is based on KVM. Originally we were working on a backup API to expose that. The teams (HPE and Veeam) decided to utilize the existing libvirt API which gives Veeam access to interact with the HVM CBT. So, HPE didn't write a new API for Veeam to use but there is some development on the HPE side to make the communication with the API and there's work on the Veeam side (Gostev, glad to hear it is looking promising). Since Veeam is using the existing API, that should in theory speed things up. In fact, I've talked to Rickatron a couple of times over the last two weeks and he said things were looking very good as well.

And if you didn't see, for those of you with a home lab, if you want the capabilities in VM Essentials, I do have the HPE Morpheus Enterprise Community Edition that includes all the features in VM Essentials. Drop an email if you're interested and be sure to mention Veeam image-based integration. Here's the email to use: morpheus_communityedition@hpe.com.
Calvin (formerly HPEStorageGuy, now the Morpheus Godfather)!

[Gostev]

Since Veeam is using the existing API, that should in theory speed things up.

Veeam is always using an existing API however, when said existing API is a low-level API of the underlying KVM, historically speaking the integration takes much longer to create and to stabilize comparing to when a hypervisor vendor provides a purpose-built backup API.

Anyway I was just trying to set expectations in response to bheungens that this integration will not be quick and simple for us like Scale was, it will be more like Proxmox which also lacked purpose-built backup API at the time.