RANT - Serious Eff Broadcom and the new Licensing.

Posted by rune87@reddit | sysadmin | View on Reddit | 281 comments

Okay...so we all knew Broadcom was going to destroy VMWare's licensing. I was expecting maybe double....nope...its now 8x the cost for my scenario. We've always used VSphere Essentials + Kit. Met our needs and I spec'd everything on the new servers to stay within the limits. Was like $700 a year. Well....that kit has now been discontinued and replaced by a new program requiring a minimum of 96 cores licensed. That's the absolute bare minimum. And at Govt cost its a little over $5500. I for one will not be paying that extortion cost. Time to go back to Hyper-V.

Reply to Post

Reply

281 Comments

[-]

Bunkaido@reddit

What is interesting for this specific scenario is that vSphere Standard might actually be less expensive. My initial calculations suggest that if you are running less than 72 cores across a cluster of 3 hosts then Essentials Plus will be more expensive than vSphere Standard. I haven't seen actual quotes to prove that yet... but according to the MSRP price lists, that will be the case.

Reply

[-]

ericneo3@reddit

> Time to go back to Hyper-V. Time to take a serious look at all the options. * LibVirt / Virtio / Virt-Manager * Proxmox * Copilot / 45 Drives Copilot for ZFS support * Hyper-V * Xen / xcp-ng Your choice will come down to performance and security. If you are fine running a web server on your virtual machine host then Proxmox, xcp-ng and Copilot are good options or you can go as bare bones as possible with a CLI server edition and LibVirt / Virtio. * If you are managing a few servers and resources look at Proxmox, it's really good. * If you are bundling many hosts and resources you should look at Xen / xcp-ng.

Reply

[-]

broknbottle@reddit

Missing LXD

Reply

[-]

ericneo3@reddit

Sorry man but no, you cannot run Windows on LXC or LXD.

Reply

[-]

Bogus1989@reddit

Well proxmox here we go! Anyone ever run it at a massive scale?

Reply

[-]

lusid1@reddit

Nobody has. The corosync used on the backend will crater long before you get there.

Reply

[-]

kingtrollbrajfs@reddit

I used to work for Broadcom. They have a policy of “we only want the big spenders as customers”. I guess that works out for them with Apple and Google.

Reply

[-]

rune87@reddit (OP)

Better have an attitude change if they want to pay off or make money on their 70 Billion purchase. Musk buying Twitter immediately comes to mind as a prime example of destroying a purchase.

Reply

[-]

kingtrollbrajfs@reddit

Look at their stock price, going back about 5 years. It’s almost 6X. I don’t agree with what they do, but they are pretty good at it - so far.

Reply

[-]

rune87@reddit (OP)

Who..Broadcom? Their stock has only jumped in the last 6 months because of the AI chip speculation and that they will profit from it. Not that they have profitted to date. If the AI craze settles down, their stock bubble will burst. Current price has nothing to do with effectively managing a company. Otherwise its been relatively flat the remaining 4.5 years. VMWare stock doubled, but that would be expected given their growth and push under Dell. Broadcom has killed Ruckus, Brocade, CA, Symantec...there is a pattern here...

Reply

[-]

Known-Amphibian-3353@reddit

Broadcom is up 2100% in the past 10 years, AI bump only happened in the past 9 months.

Reply

[-]

mrbiggbrain@reddit

>Otherwise its been relatively flat the remaining 4.5 years. What the hell are you talking about. 2018-2019 was around 25%, with similar results year over year since. There was a drop in 2022, but that has been corrected with the current rally and surpassed previous earnings. Definitely FAR from flat and a really good stock overall compared to indexes.

Reply

[-]

SmokingCrop-@reddit

https://finance.yahoo.com/quote/AVGO/ Set to 5 year or Max, they are doing really well tbh..

Reply

[-]

kingtrollbrajfs@reddit

I have had their stock for 5-6 years, and I bought at $200-ish/share. It’s close to $1,200/ share now. The AI hype is just extra gravy.

Reply

[-]

thrwaway75132@reddit

Broadcom doesn’t want to get involved with Essentials+ customers. You open a couple of support tickets and you cost more than they made on you. VMware makes like 80% of the revenue from their top 1500 - 2000 customers and VCPP partners. Broadcom will increase revenue from the 1500 largest customers by offering them more software capability for a little more money. So vSphere + Aria customers become VCF customers and Ent+ customers become vSphere Foundations customers. VMware will make more revenue from those customers, but those customers get a lot more in return. If you weren’t buying at least Ent+ then Broadcom wants to send you to OEM Channel where they don’t take the support call. Essentials customers may not like it, but Broadcom is going after the whales and will deliver them good value for their money, more focused engineering and QE on features customers actually use, and better support fir the whales at the expense of essentials customers.

Reply

[-]

pseudoanon@reddit

Isn't that self defeating in the long term? If small clients are abandoned, someone else will pick them up.

Reply

[-]

Jaybone512@reddit

And then, in a few years, if/when those new solutions get big enough, guess who will buy them.

Reply

[-]

gotoline1@reddit

Not necessarily though. At my place we are so invested in it I don't think we could pivot if we wanted to. We even have a VM engineer on staff for two years to help us. Next round of contracts in like 10 years .... Maybe. But at that point who knows what technology in this space will be like at that point. And if they double down on helping us and getting good service...why not go with them again, as a whale.

Reply

[-]

frygod@reddit

Same here. We're not huge by any means, but are around 100 hosts. Vendor support for several apps also pretty much requires vmware or hyperV (and hyperV is just a no-go.) We'll probably at least lab some alternatives though.

Reply

[-]

traydee09@reddit

Yup, broadcom knows they have enough people “locked in”, that they will squeeze every dime out of them.

Reply

[-]

traydee09@reddit

Broadcom isnt in vmware for the long term. They are interested in revenue extraction. They will basically squeeze it for all its worth until its profitability drops, then they will spin it back out in an IPO or sell it.

Reply

[-]

thrwaway75132@reddit

If VMware loses that 20% of their business, but they increase the revenue from their top 1500 by 40% while cutting their Selling and General Administrative Expenses by 50% who won? The cost to sell to and support the little customers isn’t enough to make it worthwhile. How many AE / SE pairs does it take to sell $1M of essentials to 1000 customers? How many AE / SE pairs does it take to sell $100M of VCF to a top 5 global bank? It’s the same amount of manpower for two vastly different outcomes.

Reply

[-]

GhostDan@reddit

Hyper-V has entered the chat

Reply

[-]

redvelvet92@reddit

I doubt Apple or Google use VMware.

Reply

[-]

kingtrollbrajfs@reddit

Apple was a huge VMware customer before cloud computing platforms really took hold (2010-2011). Siri used to run on thousands of vSphere hosts. Broadcom doesn’t need to sell their software products to every large customer, but by placing certain hardware and certain software into packages, they can keep the money flowing. They are diversifying. This is especially important in a post-Covid era when (traditional hardware) manufacturers got bitten, labor was hard to come by, supply chains all failed, and shipping was outrageously expensive. Software, on the other hand, is almost “free”. Sure there is the acquisition cost and some salaries, but AVGO will just close all the offices, fire half the staff, and continue to reap the rewards of VMware’s prior success. It’s very difficult, slow, and expensive to get off of these infrastructure products that Broadcom targets. Have you ever tried to just “quit” Oracle or Microsoft? It’s nearly impossible. Perfect, if you’re Oracle or Microsoft. ..or Broadcom.

Reply

[-]

a1-vergeio@reddit

I was supporting Apple with VMware back in 2019 with my storage platform at the time but they were also moving a lot of their workloads to K8s, at least the team I was supporting

Reply

[-]

TequilaCamper@reddit

Curious what OS Apple was running in their VM's? A Linux variant?

Reply

[-]

frygod@reddit

I would assume a BSD variant like all of their products?

Reply

[-]

akp55@reddit

If they are moving the workloads to k8s then it's Linux......

Reply

[-]

AnnyuiN@reddit

I would assume Linux. Afaik BSD is only used in their devices and their devices are generally considered "prosumer"

Reply

[-]

heisenbergerwcheese@reddit

So... they're not a VMware customer and havent been for over a decade...

Reply

[-]

trisanachandler@reddit

2019 isn't that long ago.

Reply

[-]

heisenbergerwcheese@reddit

So... theyre *STILL* not a customer then?!?

Reply

[-]

kingtrollbrajfs@reddit

I’m just speaking from when I supported Apple and worked at VMware, and then later, Broadcom.

Reply

[-]

admlshake@reddit

They were until some what recently, they've scaled back over the past decade, massively. It wouldn't surprise me if they still had some instances running somewhere though.

Reply

[-]

FuriousRageSE@reddit

>labor was hard to come by, Not so hard to find labor.. Just keep raising the salady until you have a full team, and then you know what bare minimum is required to find labor.

Reply

[-]

kingtrollbrajfs@reddit

I'm writing about the manufacturing of semiconductors in Asia. This would have been during heavy COVID-19 restrictions, when it was not possible to operate with a full shift.

Reply

[-]

defiant103@reddit

Both Apple and Google were VMware shops as of 2019. Believe they are still customers.

Reply

[-]

I_Survived_Sekiro@reddit

It’s harder to name a company of their size that DOESNT have VMware in some capacity.

Reply

[-]

redvelvet92@reddit

That isn’t true, the last 3 companies I’ve worked for were not on VMWare

Reply

[-]

MundaneFinish@reddit

Google Cloud VMware Engine.

Reply

[-]

AnnyuiN@reddit

Azure VMWare Service as well

Reply

[-]

joey0live@reddit

I can tell you, my company, who employees OVER 15k at higher education.. we just received an Email that they decided not to renew VMware license for server-side; client side, still okay!

Reply

[-]

dustojnikhummer@reddit

I mean they said that out loud, didn't they?

Reply

[-]

kingtrollbrajfs@reddit

Yes, straight from Hock Tan himself.

Reply

[-]

TheTomCorp@reddit

NVIDIA has entered the chat: Nvidia purchased Bright Cluster Manager and integrated it with another product so it couldn't be purchased solo. The price went from $30k per year license + support to $1.2MM for the licenses and $320k per year for support.

Reply

[-]

SnavlerAce@reddit

Fellow escapee!

Reply

[-]

cjcox4@reddit

As one living under "the truth" for a bit, we've know about the high price of VMware for many many years. Red Hat stopped doing virtualization, leaving you with expensive Nutanix, Xen Server or on the cheap, XCP-ng, as alternatives. Pretty bleak. (if you don't go the Hyper-V route, but maybe that's the lesser of all listed?)

Reply

[-]

Technical-Load9196@reddit

Native AWS Virtualization..... elastic container based App virtualization. no question its going to be the next huge leap. AWS already has a huge framework and we are retooling to make that lateral move FROM VMC on AWS. Initial numbers crunched shows double the performance @ almost half the cost.

Reply

[-]

cjcox4@reddit

Cloud is another possibility (but a change). Just realize that "handling" of reliability overall, as strange as it sounds, is still very much in "your hands". Just something to keep in mind.

Reply

[-]

Technical-Load9196@reddit

True, but with a small amount of DR built into your Maintnance plan for Cloud Based anything, You can easily find the native cloud tools to automate most everything and STILL save money overall. Our management load has decreased by 2 whole mid-teir IT employees and one help desker... Real-talk

Reply

[-]

cjcox4@reddit

Merely dispelling some "common myths". "Cloud" does "go down", and you have to figure out how you're going to handle that.

Reply

[-]

HTX-713@reddit

Don't forget Proxmox. Seriously it's the best replacement

Reply

[-]

djgizmo@reddit

Meh. Until Proxmox has a Vcenter like orchestrator with support with multiple clusters, it’ll never have a good foothold in an enterprise. Sure, maybe for single sites SMB, but not in any org I’ve worked for in the past 7 years.

Reply

[-]

Bogus1989@reddit

Damn i cant believe they havent jumped on creating an equivalent

Reply

[-]

djgizmo@reddit

Supposedly next major version should have it… but who knows.

Reply

[-]

joey0live@reddit

I would suspect yes, Proxmox devs and such is probably eyeing this subreddit and all they see is $$$ now.. if they listen to the Enterprise world.

Reply

[-]

djgizmo@reddit

They do. For the longest time though, VMware has had a stranglehold on every enterprise because their pricing was competitive for the forever license and Robo licensing… now those are going away, I hope Proxmox can hire up and take advantage of the market vacuum.

Reply

[-]

TheTomCorp@reddit

Maybe they can hire a bunch of former VMWare engineers to help build it!

Reply

[-]

djgizmo@reddit

Maybe. From what I’ve read though, it’s not the dev teams that got cut.

Reply

[-]

Bogus1989@reddit

Well hell…i was gonna move my homelab to esxi 8, now that all my hardware is compatible. Think ill keep it where im at, and maybe setup nested proxmox and give it a whirl. Might setup a nested proxmox at work too

Reply

[-]

HoustonBOFH@reddit

Not sure what you need for it that you can't get with some other monitoring... It is all web pages for management, and a monitoring dashboard can give you the at a glace status. And if they do not have more functionality built in within the next 6 months, I will be shocked. The focus for them has been the small and medium businesses, and no one expected Broadcom to mess it up this bad, this fast.

Reply

[-]

djgizmo@reddit

Lulz. Vcenter isn’t used for monitoring. It’s a host and vm orchestration tool. Promox works in a single cluster infrastructure. There are many enterprises that are REQUIRED to use multiple clusters, usually at different sites. Vcenter allows one to be able to move VMs between any host (same or different cluster) and move it live from within a gui. I’ve literally built a new DC from a template then xferred it to a remote site (1100 miles away) while it’s live. Right now the only other products that are similar are Nutanix or XCP-NG.

Reply

[-]

Fwiler@reddit

as does hyper-v

Reply

[-]

djgizmo@reddit

Yep. Missed HyperV. Its never on my radar :(

Reply

[-]

HoustonBOFH@reddit

You did hit the one gap that is not there yet, inter-cluster migration. Bit of a niche application, but I get it if you need it. Hopefully there will be more options coming and you are not stuck in the extortion squeeze.

Reply

[-]

unkleknown@reddit

Promox looks good for the small SMB sites. The essentials licensing doesn't affect the larger sites anyway so your point is moot We have 600ish SMB customers we support. Those with on-premise hosts are on Essentials and no vSphere. Just an ESXi host. Host monitoring is performed via our RMM platform. vSphere would be nice for proactive alerting, but managing 600 of those to keep them up to date doesn't make sense. With a single host, there is no need for vMotion or replication.

Reply

[-]

GeneMoody-Action1@reddit

I have never tried, but I do believe Veeam One community supports esxi directly, sans vcenter? [https://www.veeam.com/virtual-server-management-one-free.html](https://www.veeam.com/virtual-server-management-one-free.html) The requirements to not say specifically, it just lists it along side vsphere, so may be worth looking into. [https://www.veeam.com/backup-replication-system-requirements.html](https://www.veeam.com/backup-replication-system-requirements.html) They just state here they do not support "Free versions" of Esxi. [https://helpcenter.veeam.com/docs/one/deployment/supported\_platforms.html?ver=120](https://helpcenter.veeam.com/docs/one/deployment/supported_platforms.html?ver=120)

Reply

[-]

unkleknown@reddit

We have never had any issues with Veeam with ESXi without vCenter. Only with Hyper-V.

Reply

[-]

GeneMoody-Action1@reddit

Then I would assume you could leverage the community edition of Veeam one for the more advanced proactive alerting? I love my Veeam one, but it is also an enterprise setup with full vSphere. So never tried it on that level. Worth looking into maybe?  If you do, I would be interested to know it worked. I do not see myself needing it anytime soon, but always nice to know in case one day I do!

Reply

[-]

djgizmo@reddit

600 individual small customers does not equal an enterprise. Proxmox works fine for small customers with one site. It does not work well for those that have multiple sites and have the need to move VMs to different sites. Your point is both moot and mute.

Reply

[-]

unkleknown@reddit

In the original post, they are talking about Essentials license. Thus they are not a large enterprise. I never said anything about 600 customers being an enterprise. I've worked in enterprise in the past and see the point in using the appropriate licensing; I was replicating off-site and would use vMotion to shift guests for hardware maint. The OP obviously doesn't have more than 3 hosts with Essentials licensing. What do you think they should use, vSphere standard? It's much more costly, like 10x more.

Reply

[-]

TheDawiWhisperer@reddit

that's my attitude to it...proxmox has it's place (arguably that place is homelabs lol) but it's not fit for enterprise at the minute

Reply

[-]

HoustonBOFH@reddit

I have been seeing it in larger businesses, and education for a while now. This will just push it more. Yes, it lacks some large enterprise features. But not for long!

Reply

[-]

djgizmo@reddit

Not yet ;) Hopefully this VMware change lites a fire under both Proxmox and XCP-NG.

Reply

[-]

Key-Calligrapher-209@reddit

Did I miss something? Red hat still offers hci and open shift, don't they?

Reply

[-]

TheTomCorp@reddit

Yea they dumped RHEV (based on oVirt) which was a solid project. Openshift Virtualization is their new virtualization platform, they also offer Openstack (build on top of Openshift). It's all more complicated than it needs to be, and isn't cheap. Oracle still has their virtualization platform built on oVirt. If Oracle is the last hope for oVirt... What a sorry state of affairs, it really is a race to the bottom.

Reply

[-]

Iamnotapotate@reddit

Oracle virtualization is awful. It runs well enough once it's up and going, but doing any kind of management of it is like sticking your hand in a bucket of razor blades.

Reply

[-]

lightmatter501@reddit

Openstack is a toolkit to build a private cloud. Past a certain scale it’s worth it, but it’s not worth it for a few hosts.

Reply

[-]

luckenbach@reddit

Openshift offers KubeVirt as an option but I feel like you have to design with the known limitations for this to be viable for you

Reply

[-]

SquizzOC@reddit

Hey at least you can get a quote. I keep asking for a renewal for a client: Please quote me instance 1, instance 2 and instance 3 for a 1 year renewal, only thing the client is renewing all other license's are done. "We can only renew instance 5 at this time" They have Broadcom employees stepping that know ZERO about VMware. Absolutely nothing.

Reply

[-]

Technical-Load9196@reddit

*" ....They have Broadcom employees stepping that know ZERO about VMware. Absolutely nothing...."* And \*THIS\* is what will kill them if they arent careful. With VMCloud and so many other VMWare-to-native solutions just having been fully birthed.... the Hands on and Ability to answer the simplest questions will be critical.... And I have a hunch this is what is going to hurt them.... bad.

Reply

[-]

SquizzOC@reddit

I had a call with one of my largest clients today and they had a 10 year tenured engineer on the line to answer technical questions, but the sales rep was clearly shuffling through a power point to answer questions sadly :(

Reply

[-]

FriskyDuck@reddit

I'm kind of shocked our renewal was processed within 2 days after submitting it to our VAR last week.

Reply

[-]

epsiblivion@reddit

Grandfathered socket license or forced core licensing?

Reply

[-]

epsiblivion@reddit

I need to hurry up and ask for quotes then before the door shuts.

Reply

[-]

FriskyDuck@reddit

Grandfathered perpetual socket for another year. At least we can budget properly for the next FY renewal.

Reply

[-]

Aggravatingrer0@reddit

Did I miss something? Red hat still offers hci and open shift, don't they!

Reply

[-]

Nick85er@reddit

yeah they allowed us a 1 year courtesy renewal of our "perpetual" license as well this month - the audacity of these sales people to maintain a straight face and stick a hand out for 5 figures.....

Reply

[-]

NISMO1968@reddit

It’s shocking situation actually…

Reply

[-]

rezadential@reddit

What does one need to move to a production level hyper V environment? Just the baremetal hypervisor? Last time I looked, it seemed to lack a lot of features and you needed something like SCVMM to manage it similar to how one would use vCenter to manage vSphere environment

Reply

[-]

HTX-713@reddit

Or just use Proxmox. For some reason everyone is just shitting on it when it is literally the best replacement. Other open source alternatives have been discussed that aren't close to being as feature rich as Proxmox, but every time Proxmox is mentioned everyone complains about "lack of support", when they do have support that can be purchased. Honestly I think people are afraid to learn something new, even though it's just a GUI over Debian and KVM.

Reply

[-]

pbrutsche@reddit

The cost of dotting-the-i and crossing-your-t for your entire stack is very high. There are a lot of hypervisor options out there.... if all you are looking at is a hypervisor. Backups are a major issue, and a lot of the backup "solutions" for a hypervisor that's not one of the top 3 (VMware, Hyper-V, Nutanix) are amateurish toys or require you to seriously revamp your tooling and thus drastically increase costs. One big area where that shows up is the change from hypervisor-based backups to agent-based backups. The cost increase to the VMware vSphere subscription is insigificant compared to the cost to change Veeam licensing from socket-based to VUL. I haven't even touched on the point of third party integrations. It's going to be an unpopular opinion, but it's no different than trying to compare a basic SPI firewall like pfSense to a FortiGate or Palo Alto. Absolutely no comparison. Trying to say that Proxmox is in the same league as VMware or Nutanix is just as laughable as saying that pfSense is in the same league as a FortiGate or Palo Alto.

Reply

[-]

datarapan@reddit

Evaluated alternatives due to this solution issues in VMware backup. Switched to nakivo for per-VM licensing a year ago. Initially skeptical about snapshot and recovery, but everything integrated seamlessly, performed reliably. Transition costs justified by substantial license fee savings.

Reply

[-]

Soggy-Camera1270@reddit

While I love Proxmox, you are forgetting that "support" doesn't always mean vendor support. Often it refers to lack of third party support and/or integration. It's not so simple to replace a full suite of tooling just because you want to move to a different hypervisor - particularly when you are talking about hundreds of hosts. In small environments, yes, that would be pretty easy and a solid option.

Reply

[-]

roiki11@reddit

This. While proxmox is nice, it has nowhere near the integrations vsphere has.

Reply

[-]

Maraxius1@reddit

Are there any added features or support to justify the increase, or is it just them slapping their brand on the existing product and jacking up the price? If the latter, they're just going to end up forcing the majority if their customers into competing products...

Reply

[-]

King_Of_The_Cold@reddit

ill give you 3 guesses

Reply

[-]

Maraxius1@reddit

https://preview.redd.it/txl12p2g7ubc1.jpeg?width=750&format=pjpg&auto=webp&s=fe90d6ead7880552fc1ac4b46c38a526a9aca123

Reply

[-]

Berg0@reddit

we're seriously considering rolling to Hyper-V for our IAAS hosting at the colo. Almost all of our clients with on-prem virtualization needs are on Hyper-V already anyway, with a few stragglers still on vSphere, that will either shift to cloud or get Hyper-V at next forklift upgrade.

Reply

[-]

gotoline1@reddit

This might be a dumb question, but would KVM work on this scenario at all? I've used it for small applications before and worked well. I haven't tried hyperV since like 08 or something like that.

Reply

[-]

sobolrocket@reddit

>Starwinds HCA KVM wrapped in a more usable interface may work for you. Try ProxMox.

Reply

[-]

mr_ballchin@reddit

That is the point. We found that for small 2-node environments, SDS solutions like Starwinds HCA, based on Hyper-V and provides iSCSI block storage for HA, are a decent option. The support is also great, which covers the lack of solid Microsoft support.

Reply

[-]

546875674c6966650d0a@reddit

Broadcomm's hq in the bay area had a legit Lamborghini parking only area. This is how.

Reply

[-]

unkleknown@reddit

NNNNOooooo, don't go to Hyper-V. I've been playing with Promox and I'm pretty happy with it. We are in the same boat with our customers. Historically, in our experience using Veeam, Hyper-V sees too many backup failures due to snapshot issues.

Reply

[-]

MFKDGAF@reddit

Why version of hyper-v? I use to seen tons of failures too but that was on 2012 R2. I thought proxmox didn’t handle windows VMs that well but was excellent for Linux VMs?

Reply

[-]

jess-sch@reddit

>I thought proxmox didn’t handle windows VMs that well but was excellent for Linux VMs? I have no idea where that claim comes from. Windows VMs work just fine, as long as you install the drivers from Red Hat's virtio-win ISO.

Reply

[-]

unkleknown@reddit

Our bad experiences with Hyper-V are historical, so 2012 R2 and 2008 R2. As far as Windows, I have Win 11 and Server 2022 running without any issue on Promox. Had to dig with it a little to get the Virt Viewer running but now it seems to work well. Also have a virtual Sophos firewall running which is Linux based. Using an old Dell T320 with a 10 core CPU, 48G RAM and 8 SSDs. It's surprisingly fast. If I have HV running on this, it wouldn't be anywhere this fast from previous experience using 2019.

Reply

[-]

Doso777@reddit

> Hyper-V sees too many backup failures due to snapshot issues. They are called checkpoints for almost a decade now. We are all-in Hyper-V for over a decade. I can remember maybe a handful of problems with backups. Most of them fixed by restarting the hypervisor (Windows Server 2012R2) or with a quick "delete old snapshot" powershell command (Windows Server 2019).

Reply

[-]

Dushenka@reddit

We went with VMWare on our last server revision because it was the easiest and most affordable option. But that was 5 years ago and the competition caught up, big time.

Reply

[-]

Model_M_Typist@reddit

I finally got approval and bought new hardware for a 3 node cluster and a SAN, with VCenter licensing. So excited to move off Hyper-V. Government pace is slow as fuck. I heard about the acquisition before receiving hardware. I've gotten the hardware now but I'm just kinda over it

Reply

[-]

ChumpyCarvings@reddit

Proxmox is going to improve due to this right? Like surely the industry is going to work together to get this done in the next 2 years?

Reply

[-]

NISMO1968@reddit

I doubt it! They are around since 2008 and had lots of time to do their homework.

Reply

[-]

ChumpyCarvings@reddit

Proxmox is already very good, I am talking more about larger features which VMWare has.

Reply

[-]

NISMO1968@reddit

>**Proxmox is already very good**, Define "very good". >I am talking more about **larger features which VMWare has**. So do I. It's backup comparable to CommVault or Veeam feature-wise.

Reply

[-]

ChumpyCarvings@reddit

Proxmox is reliable, pretty well known, extremely well priced, can do a variety of functions which extend beyond homelab stuff. As for the features, yes I hear the backups need improvement.

Reply

[-]

jess-sch@reddit

The main complaint about backups seems to be "it doesn't support Veeam". Which, yes, but would someone be willing to explain to me how the backups via Proxmox Backup Server are insufficient?

Reply

[-]

caa_admin@reddit

What does improvement mean in context. Proxmox slays ESX in many ways. No, I might not use PVE in an HE or clustered manner at this time like I would ESX. Not all admins need HE or clustering for their needs either. But, give it time. Proxmox will 'catch up' to whatever some ESX admins consider deal-breakers.

Reply

[-]

ChumpyCarvings@reddit

Sorry let me rephrase, I know Proxmox is great. I'm hoping in time, proxmox will become even greater with whatever features it lacks from VMWare added in time.

Reply

[-]

BoilingJD@reddit

why does everyone forget Proxmox Exists? It is cross compatible with everything. Easy to learn, great community, paid support that is not utter dogshit if you want it. and Linux KVM is arguably the ultimate hypervisor. If you want full white gloves support, then try Verge.io. Their product is built on Linux with their own special sauce for clustering. i've got 3 proxmox clusters and one Verge.io cluster, collectively 100 VMs or so, doing HPC compute and GPU passthrough. Between Prox for experimental hacky techops stuff and Verge for full on private cloud experience with great support, I genuinely don't see how can Hyper-V or whatever Citrix is selling can even come close to this.

Reply

[-]

Stratbasher_@reddit

Proxmox is difficult with shared storage. It's damn near impossible to get thin provisioned disks and snapshots in a cluster without paying for specialized hardware or software. I debated switching my home lab over to it but VMware is still used in enterprise and is easier to get working if your hardware supports it.

Reply

[-]

BoilingJD@reddit

Proxmox supports SMB, NFS, iscsi connectivity, and ZFS and Ceph out of the box. How is that 'difficult with shared storage' ?! literally covers any possible protocol you would ever use. And if that's not enough you can always go into the debian backend and install whatever native client you want. Quobyte, GPFS or whatever tickles your fancy.

Reply

[-]

roiki11@reddit

No nvmeof. Would have to go back to fc. You don't run databases off iscsi.

Reply

[-]

BoilingJD@reddit

how much storage do you need for your DB ? Why not use Nvme backed Ceph ? why not use SMB or NFS? what kind of performance do you need to achieve ?

Reply

[-]

roiki11@reddit

"A lot". And it seems you don't know much about databases. They care a lot about latency and iops. And neither ceph, smb or nfs are suitable for any serious workloads. It's either fc, roce or direct storage.

Reply

[-]

BoilingJD@reddit

If you have databases that need stupendous ammount of Iops then why are you not using Weka, Quobyte or GPFS like all sane people ? Hypervisor's shared storage is for allowing HA and roaming VMs across compute nodes. Which is not super resource intensive, All actual data should live in a dedicated structure for it.

Reply

[-]

roiki11@reddit

Because none of those is for databases? And all of those can't hold a candle to a pure array, they still use normal network protocols which will always have more overhead than fc or roce. The better alternative would be infiniband. We use pure arrays like all sane people.

Reply

[-]

BoilingJD@reddit

So if you are running Pure anyway, what difference does it make if your VMs are in ESXI or KVM ?

Reply

[-]

roiki11@reddit

All the management that's built into vsphere. Vvols, the vsphere plugin, datastore clustering. Sure, you can run it on a normal linux like any storage array. And ESXi is not different in this regard. It's the features in vcenter that make it really powerful.

Reply

[-]

BoilingJD@reddit

all of these features either solve a problem that is a problem of Vmware's own creation or a problem that can be solved another way.

Reply

[-]

roiki11@reddit

Lol what.

Reply

[-]

Fwiler@reddit

It doesn't matter. You talk as if you know everyone's compute and storage needs. Why are you even commenting? Apparently you can only use fc, roce, or direct storage if you know anything about databases? lol.

Reply

[-]

Fatel28@reddit

Proxmox lends itself more to distributed storage for HA over a SAN or some other shared storage device. We just setup a 5 node cluster for our IaaS offering, and so far HA with Ceph has been excellent.

Reply

[-]

Stratbasher_@reddit

What I mean is that I want to run VMs on a cluster of proxmox hosts pointed to my SAN. There's no native clustering filesystem like CVFS or VMFS, so you've gotta figure it out on your own. Snapshots become LUN snapshots instead of per-VM snapshots at the hypervisor. Otherwise you have to use a filesystem in the VM to provide snapshots. And thin provisioning would normally be done with lvm-thin but you can't use that on shared storage IIRC. So you end up shoving ceph into there and it's just a mess.

Reply

[-]

Fatel28@reddit

If you have an ISCSI capable San you can put a ZFS pool on the ISCSI filesystem and take advantage of snapshotting and error checking etc. It's actually even documented in the proxmox forums. Putting Ceph on a San doesn't make sense, that's not the point of Ceph.

Reply

[-]

teeweehoo@reddit

I thought ZFS over ISCSI required a server running zfs and ISCSI. Then proxmox uses SSH to access the server to create new LUNs. I'm not aware of a way to use ZFS on an ISCSI share from a SAN, besides making a LUN per server.

Reply

[-]

Fatel28@reddit

ISCSI just presents as a volume to the OS, and you can do ZFS on a single volume. That's more what I'm referring to.

Reply

[-]

teeweehoo@reddit

Right but ZFS itself isn't a clustered file system, so you lose all the advantages of shared storage. You could always map the ZFS volume to another host, but that's quite manual.

Reply

[-]

ericneo3@reddit

> ZFS itself isn't a clustered file system No, but you can use Proxmox / TrueNAS to manage ZFS then you pass up the storage to Hyper-V via virtio-scsi or iscsi and have Hyper-V create shared clustered storage while also allowing you the benefits of one or both ZFS & KVM block caching.

Reply

[-]

Fatel28@reddit

That is true. I'm not disagreeing with your point that proxmox isn't the best if your use case requires a SAN. Like I mentioned, proxmox leans heavily into distributed filesystem, not shared.

Reply

[-]

_BoNgRiPPeR_420@reddit

I have no doubt it's good for virtualizing things, but what backup solution does it have that's comparable to the likes of Veeam in terms of features and performance? Veeam supports Red Hat virtualization which is also KVM, but I don't think they officially support Proxmox.

Reply

[-]

BoilingJD@reddit

1. You can back up to any network storage you want 2. you can do ZFS snapshots to a ZFS storage 3. You can use Proxmox Backup Server which fully manages delta backups for you. What else would you possibly want ?

Reply

[-]

ericneo3@reddit

> What else would you possibly want ? * Mountable file-level * VM disk backups The quicker and more ways I have to recover something or some part the better.

Reply

[-]

caa_admin@reddit

> Mountable file-level I've not tried but I believe you can. > VM disk backups Not sure what you mean. It's built into proxmox.

Reply

[-]

_BoNgRiPPeR_420@reddit

File-level restores are a big one for us. I'm not sure if Proxmox backup server handles those. We also use Veeam's backup and replication to automatically replicate VMs to a DR site. It does application-level backups for SQL Server and a bunch of other stuff too. Integrations to use scale-out repo to copy data to Amazon S3 storage in the cloud. I'd be hesitant to take a step backwards if these features aren't available.

Reply

[-]

BoilingJD@reddit

If you use Verge.io, they have built in geo-replication feature that works equally for individual vms or entire environment complete with virtual network config etc... Proxmox can also do ZFS delta replication to another remote proxmox cluster. For file level restore, verge has very powerful snapshot and deduplication engine built into it's vSAN. You can restore a vm from snapshot into a separate virtual network and pull out whatever you need. Alternatively, there is a million different ways to mount a raw disk image to pull out specific files from it.

Reply

[-]

ericneo3@reddit

> Linux KVM is arguably the ultimate hypervisor. Time, experience and equipment to test with. From my experience the biggest problem was not having the equipment to test, compare and learn. Once I started home lab testing I learned how it all worked and saw just how much better it was but that is a lot harder to do with in production equipment. > Between Prox for experimental hacky Not the words people want to hear with a hypervisor. The biggest problem to Proxmox adoption is going to be busted Linux updates/upgrades that break the host, Proxmox is no exception to this Linux problem. Immutable distros will fix this, each update/upgrade will produce a new Grub boot entry and if an dev, app or distro push something completely busted, the solution will be a quick restart of the host and load to the previous working entry from Grub.

Reply

[-]

BoilingJD@reddit

We're running proxmox in a very unstable environment and it runs rock solid. Obviously one needs to know some degree of general linux sysadmin to know what to do and not to do, but that 's sort of a given for a sysadmin in 2023.. However, we also run a Verge.io cluster as our primary client-facing 'private cloud', and it's rock solid. Super easy to use and the support is the best support I've ever experienced from any vendor, ever.

Reply

[-]

hurkwurk@reddit

no disrespect to proxmox, but especially in government, you pick a market leader, or you spend the rest of known time justifying why you didnt, and worse, if anything serious actually went wrong, its likely your ass or your career track. the best solution often bows to the political whim.

Reply

[-]

rune87@reddit (OP)

Pretty much. Brings the old saying "No one ever got fired for using Cisco" to mind.

Reply

[-]

tsmith-co@reddit

Welp. I guess I’m officially old now.

Reply

[-]

tmpntls1@reddit

The original saying was actually "No one ever got fired for buying IBM". Apparently we're older than we thought, Tim. If the cheap imitation of our old saying is now old...

Reply

[-]

tsmith-co@reddit

Exactly! Just think, 20 years from now it will be remembered as “no one ever got fired for buying VMware” 🤣

Reply

[-]

tmpntls1@reddit

Hopefully we're all retired and running our massive karting commune & distillery. lol

Reply

[-]

closedgate@reddit

Nobody ever got fired for hiring Joe or Tim.

Reply

[-]

TylerJurgens@reddit

That's a given!

Reply

[-]

rune87@reddit (OP)

Amen. I won't care what the latest IT crisis of the year is. "Quantum Virus Unstoppable!" "Skynet goes live"

Reply

[-]

rune87@reddit (OP)

LOL! ....Who/What is IBM? 🤣

Reply

[-]

ericneo3@reddit

> "No one ever got fired for using Cisco" to mind. Until their management saw the bill... XD

Reply

[-]

HTX-713@reddit

The fact that they balk at $5k increased licensing costs tells me they are small government. Proxmox is super easy to set up and is rock solid. They would save money by moving to it.

Reply

[-]

NameEnvironmental358@reddit

Until they can provide 24/7 support, most will not be comfortable using them for production. You probably won’t ever need support, but management will say otherwise.

Reply

[-]

libach81@reddit

They previously had a text on their site stating support was only provided in business hours in CET timezone, but I can't see that anymore. So it that still the thing? If yes, there are Proxmox partners offering 24/7 support for the product and since it's Linux with KVM, I'd say a capable Linux shop will come quite far in troubleshooting issues. And for those issues that requires developers to patch (and where it's actually the specific Proxmox code), not even Vmware is going to get that patch out the door in those few hours it takes for a Proxmox dev to wake up somewhere in the world.

Reply

[-]

valdecircarvalho@reddit

Until you need. Manager’s neck is on the line

Reply

[-]

DeesoSaeed@reddit

I believe that with a good bunch of SMB moving to Proxmox at some point they will have the financial leverage to close gaps in their support offerings and features. It'll take some time, though.

Reply

[-]

Wonderful_Guitar2178@reddit

In my particular case I have a multi-vendor hardware environment with a lot of linux, so if part of my environment gets pricey we will just move the workloads elsewhere. Broadcom has always been a janky company in terms of support for their network and fc hardware. If the juice isn’t worth the squeeze I’m flexible enough to go elsewhere, I’ve never been locked in with a particular hardware or cloud provider

Reply

[-]

kjstech@reddit

I had a quote prepared a few months ago, expires this week with my reseller. I was able to execute it yesterday, thank god. It was basically to make our DR setup match PROD. Was the vSphere essentials and 4 hosts, now we’re full vSphere and 8 hosts for DR. I just got my license confirmation emails today. I’m so lucky I didn’t wait till the quote expired and had it prepared late October. On these perpetual licenses were good till 2026 so have some time to think if hyper-v, xcp-ng or proxmox are alternatives. Thing is we’re using SRM and array to array replication, iSCSI to pure Flasharrays. All ESXi hosts are Dell servers. New m7000 chassis (8 blade) and two FX2 chassis (4 blades each).

Reply

[-]

cpbpilot@reddit

You could always try proxmox

Reply

[-]

Magyarharcos@reddit

I was kind of expecting this, but man, i am so sick of publicly traded companies pushing for profit at all costs I bet they are going to make it be a subscription service forever, like adobe did, with server authentication just to look out the filthy pirates. Its going to work about as effectively as adobe's efforts too, i hope.

Reply

[-]

yParticle@reddit

These subscription rates for on prem licensing are ridiculous. Might as well go with AWS or the like and not pay all the hardware costs. Miss the days of perpetual licensing and keeping a hypervisor across several OS upgrades.

Reply

[-]

d00ber@reddit

I've priced it out for my company because of the data we move. It would still need to go up 200x for it to cost as much as AWS for us. Obviously, if you don't have as much data in transit or need as much data storage that's not a big deal. That said, I'm considering proxmox, and general containerization. I have some pretty field related applications that are even starting to move to docker, which is shocking cause the field has been stagnant for decades.

Reply

[-]

Mysterious-Safety-65@reddit

Just curious....what applications do you have running in docker containers?

Reply

[-]

d00ber@reddit

Mostly web applications for project management, asset management, some custom web apps for automations. The things that are moving towards containerization are GIS specific applications.

Reply

[-]

SA_22C@reddit

Spoken like someone who has never actually costed out IaaS on AWS.

Reply

[-]

SpectralCoding@reddit

> Pays for 5 years worth of hardware on-premises, doesn't bat an eye > Runs $50k/mo with everything On-Demand because they're afraid of Savings Plans because they might retire two servers in the next year.

Reply

[-]

SA_22C@reddit

To whom are you referring ?

Reply

[-]

SpectralCoding@reddit

People's internal cognitive dissonance when it comes to cloud costs. They have no problem dropping $300k every 5 years on new VM cluster hardware, but then they move to the cloud and don't take advantage of Savings Plans or Reserved Instances. They complain about the price because they're paying a premium for on-demand pricing instead of paying for what they actually need, which is SPs/RIs.

Reply

[-]

SA_22C@reddit

Even with savings plans and reserved instances, IaaS is roughly 2x more expensive than collocated VMware running on Cisco UCS. I’ve done the math, and the cloud providers are ripping you off.

Reply

[-]

SpectralCoding@reddit

It's probably close to a wash if you just compare hardware to hardware for the full stack. But in my experience people typically conveniently ignore the features they don't think they need but would DEFINITELY use if it were easy or built-in. It's all good to compare "a few racks in a colo" to AWS but this gets old when you tell management: * in order to do proper firewalling we need to spend an additional $60k on firewalls * in order to do backups we need to license backup software and buy tapes * in order to get network switches for a new HPC project we have to wait 11 weeks * in order to do literally any valuable BC/DR we need to do all that in a second physical location * our hypervisor vendor got acquired they're jacking up the price 8x Me? I'll log in, spin stuff up immediately, and take the free features and upgrades I get literally daily for literally zero effort and zero incremental cost.

Reply

[-]

SA_22C@reddit

Thought of all that. Included all that. It’s still massively more expensive.

Reply

[-]

surveysaysno@reddit

They aren't listening. Lots of people have done the math, and we just get told we "don't understand" the complexities. Cloud is more expensive. If a business has a hardware footprint larger than 4 racks, the economies of scale for on-prem/co-lo are such that cloud is more expensive. The end.

Reply

[-]

mrdeadsniper@reddit

I mean, I think that should be the assumption. Companies aren't hosting on the cloud for the charitable spirit of information technology. They are selling a service, which means they are charging more than it costs them.

Reply

[-]

cs_major@reddit

From what I have seen...The people that screwed the most on cloud just pick up the VMs and drop them on the cloud instead of moving them to the native offerings.....SQL and file storage come to mind.

Reply

[-]

noobposter123@reddit

The others are screwed when their cloud provider starts screwing them and they can't move off the "native stuff" easily. Whereas those with VMs can more easily move to another cloud provider...

Reply

[-]

HTX-713@reddit

This right here. Most of the time it's because companies don't want to spend the investment to re-architect everything to take advantage of the very features that will save them the most money with cloud.

Reply

[-]

cs_major@reddit

Well and don't do it all at once. Create a list of applications or products that are low hanging fruit and easy to shift as a starting point. Don't just move every single VM at once and cross your fingers.

Reply

[-]

Usodus-3389@reddit

You are correct savings plans and RI are critical to keeping the dollar spends reasonable

Reply

[-]

placated@reddit

It can be comparable if you aren’t dumb about sizing workloads and you are willing to commit to a certain level of spend.

Reply

[-]

SA_22C@reddit

No, it really can’t. If IaaS is cheaper than your on prem costs, you’re doing on prem wrong.

Reply

[-]

yParticle@reddit

Depends on your load. Most services & EC2s are actually pretty affordable at the entry level, it's when they need to scale up that the costs get crazy. I've just used them for small operations so they're cheaper than colocating servers at our data center.

Reply

[-]

mr_white79@reddit

Another Broadcom/VMware thread, another notch in my 15 years of experience with Hyper-V finally being an advantage.

Reply

[-]

heymrdjcw@reddit

I’ve already got my first quarter filled with Hyper-V and Azure Stack builds. I’m loving it!

Reply

[-]

jamesaepp@reddit

> Azure Stack I say this genuinely...good luck.

Reply

[-]

cs_major@reddit

Azure has been pretty solid for us.....Office 365 on the other hand has been not as reliable(though recently has been way better).

Reply

[-]

jamesaepp@reddit

FWIW I (and I think the person I am responding to) are talking specifically about Azure Stack HCI which in some ways blurs the lines between Azure as a public cloud and on-premise HCI. There's no specific connection here to Office 365 (SaaS) or "Azure" more broadly apart from the brand name.

Reply

[-]

heymrdjcw@reddit

I’ve got a few dozen HCI stacks out there I work with, but thank you. It has its moments but I haven’t found it any fussier than vSAN after yet another patch.

Reply

[-]

rune87@reddit (OP)

I'm thinking when the new hardware comes in next month that will be the project.

Reply

[-]

eruffini@reddit

For $5500 I'd just get the VMware licenses updated instead of going Hyper-V. The time and effort involved alone would make that a wash. Plus even with VMM the difference between vSphere and Hyper-V is still in VMware's favor.

Reply

[-]

rune87@reddit (OP)

In my case, I have two separate clusters. So that's 11k a year in licensing now. Pain in the ass..yes, but that more than justifies putting the lifting in now over a 5 year cost cycle. Hyper-V has its limitations for sure, but between the mass exodus/firing of staff and now the licensing, I suspect VMware is soon to be a shell of itself and many of us will be looking for alternatives.

Reply

[-]

justlikeyouimagined@reddit

~0.1 FTE to not have to redo/convert everything, risk fun new problems, or deal with SCVMM? Sign me up..

Reply

[-]

rune87@reddit (OP)

I'm weighing this in the bigger picture of Broadcom screwing VMWare up like everything else they've aquired. Taking the first lifeboat off the sinking ship instead of the last one.

Reply

[-]

jungleboydotca@reddit

Is the price of Hyper-V going up soon? If not, what's the hurry? If your team isn't as equally proficient in the new environment, there's a cost there too. Is the backup solution easily migrated? Plan your exit, but the platform isn't burning *quite* yet.

Reply

[-]

rune87@reddit (OP)

I should have been clearer in the original post. It's Budget Season and I have a hardware refresh. My VMware licensing is up in 10 months. Today was getting the quotes for next years purchase.

Reply

[-]

jungleboydotca@reddit

That \*\*is\*\* a hurry-up situation then. Indeed for you, I'm sure it does feel like everything's on fire. GL with the project! 🤜🤛

Reply

[-]

Lanathell@reddit

You're getting a quote now for a purchase in 6 months? I hope the price doesn't double during that time

Reply

[-]

justlikeyouimagined@reddit

All they have to do is.. basically nothing.. and the software will still be better than Hyper-V. Unless it's your company, I would smash the easy button at the office(no need to play on hard mode), and start upskilling for what's next.

Reply

[-]

Usodus-3389@reddit

What is next?

Reply

[-]

cookerz30@reddit

It seems like Microsoft is at least pushing everyone to Azure HCI. I'm already evaluating it myself. Starwind has been one of the better solutions I've seen so far.

Reply

[-]

jungleboydotca@reddit

Runtimes as a service with automagic storage.

Reply

[-]

hunterkll@reddit

Did a vSphere to Hyper-V migration a while back (3-node VMWare to 4-node hyper-v, huge jump in capacity) .... ran the converter (5nine i think i used at the time?) - clicked source VMware, destination hyper,v selected the VM, clicked go, booted up VM on hyper-v, done.  You don't have to redo anything, and conversion is automated. Plenty of tools to do it across the board that's literally just a few clicks. Hell, our backup system can restore a VM to AWS/Azure/Hyper-V/XenServer/VMware regardless of where it was originally backed up.

Reply

[-]

doweactuallycare@reddit

> For $5500 I'd just get the VMware licenses updated instead of going Hyper-V. The time and effort involved alone would make that a wash. > > Thats how you end up being the dummy that eats $5k/year increases. The fog in boiling water thing. **It doesn't end there, you're not evaluating a single $5500. You understand that, right?**

Reply

[-]

a1-vergeio@reddit

full disclaimer, I work for VergeIO We heard the same thing from a customer today about how high VMware is rising their prices, charging rediculous rates per core, it's insane if you are looking for a great VMware HCI alternative, I would seriously check out VergeIO I have worked here for almost 2 years and it's a solid product. I have also worked wth VMware since 1.5 on ESX and since 2000s on NT, and have my VCP for way too long

Reply

[-]

caa_admin@reddit

Consider an AMA, I bet some would be interested.

Reply

[-]

drosmi@reddit

96 cores is almost a large single server these days

Reply

[-]

dizzygherkin@reddit

Anyone who uses autosys is also cursing, gone from normal annual license model to “per run” … completely blows dev jobs out of the water.

Reply

[-]

d00ber@reddit

We're evaluating Proxmox.

Reply

[-]

WhiskyIsRisky@reddit

We run Proxmox at my shop. Really depends on the use case whether it's appropriate. I can say though I don't miss VMWare. I'm most comfortable being a Linux admin and Proxmox is just a nice web UI on top of Debian. After two years running it I still haven't needed support for anything.

Reply

[-]

Brandhor@reddit

how well does proxmox works for windows guests?

Reply

[-]

No-Alfalfa1894@reddit

Just fine, although it does need some extra drivers (including during the install), but you can make a template VM, to clone off of, which saves you the hassle. https://pve.proxmox.com/wiki/Windows_10_guest_best_practices

Reply

[-]

dustojnikhummer@reddit

Well you can install without VirtIO

Reply

[-]

No-Alfalfa1894@reddit

True, however I found that VirtIO things are some of the best on proxmox.

Reply

[-]

dustojnikhummer@reddit

So do I.

Reply

[-]

WhiskyIsRisky@reddit

I've not had any issues running them. We have a few Windows guests and they install and run without any weird configuration. Maybe if I switched their graphics over to Spice I'd need to install some drivers, but otherwise they just worked.

Reply

[-]

d00ber@reddit

I figure I'd be the same. The GUI is more for my coworkers, since I also have a history as a Linux admin (and windows). I actually have proxmox administration experience from 5+ years ago where we were cloud-initing vms through automations at a startup. I pretty much just have to do a demo to my coworkers and show that it can do everything that we have now. The only blank space I have right now is cloud DR, since our current cloud service is extremely cheap and offers DR for VMware vms.

Reply

[-]

WhiskyIsRisky@reddit

What cloud service are you using? Cloud DR is a definite gap for Proxmox. We have something cobbled together with Proxmox Backup Server and rclone but it's far from ideal. I could get us back online in a DR scenario but it won't be pretty.

Reply

[-]

Brook_28@reddit

Look at scale computing

Reply

[-]

d00ber@reddit

Thanks for the suggestion, but it doesn't look like the solution for me!

Reply

[-]

Brook_28@reddit

Worth a look. No worries

Reply

[-]

jktmas@reddit

I did, and they made nutanix look cheap.

Reply

[-]

d00ber@reddit

Yeah, not to mention their website is just a bunch of marketing bullshit. It takes a lot of navigation before you start to find anything of substance.

Reply

[-]

Soggy-Camera1270@reddit

100% this. I hate vendors that hide information and pricing behind smoke and mirrors marketing and "contact us for more information" BS.

Reply

[-]

a1-vergeio@reddit

they also only scale to 8 nodes, not much scale there

Reply

[-]

a1-vergeio@reddit

they also only scale to 8 nodes, not much scale there

Reply

[-]

HouseCravenRaw@reddit

I understand Proxmox only offers support during Austrian business hours. That'd make it a non-starter for me.

Reply

[-]

FreeBeerUpgrade@reddit

I'm getting sick and tired when people will STILL throw that argument in the air while many people here have pointed that ProxMox has a partner program. And theses companies will do 24/7 support. It's like saying you can't go AD because M$ only does support during business hours. At this point it's either lazyness or ch'toupidity

Reply

[-]

d00ber@reddit

We'd have a local support as well. I also have a lot of Proxmox experience and our needs are extremely low.

Reply

[-]

nerdyviking88@reddit

There are local partners (Or more local than Austria) that can offer support. https://weehooey.com/pages/proxmox

Reply

[-]

a1-vergeio@reddit

check out VergeIO too, but its an HCI stack so you can't use external SAN/NAS

Reply

[-]

dustojnikhummer@reddit

Microsoft and Proxmox will benefit heavily from this.

Reply

[-]

MURICA69USA@reddit

I think you are absolutely correct to consider switching. What people are ignoring is this is only the first of many price increases.

Reply

[-]

doctorevil30564@reddit

I have experience with proxmox, but I haven't used their backup server setup yet. We currently use Veeam with our VMware hosts, so I need to setup a proof of concept server and example VMs to show our IT director. I think he will be receptive to the change. I already have the necessary experience for doing migrations from a VMWare VM to a proxmox VM. One of my projects at my previous job was migrating a bunch of production VMs from an older 5.x esxi host to a proxmox cluster. They used cron scripts to make archive copies of the lzo backups they ran nightly on the cluster for long term backups.

Reply

[-]

Gatt_@reddit

Just as well we're already half way through our VMWare to Hyper-V transition Suspect this will be ramping up fast in the New Year now

Reply

[-]

alhttabe@reddit

I feel like Microsoft’s next earnings report is going to mention a massive bump in Server DC SKU sales

Reply

[-]

dbh2@reddit

Time to buy calls

Reply

[-]

everettmarm@reddit

Take this as you will: if it were me, I would tell the business to increase my budget due to the vendor’s increase in licensing. Then I would use the additional budget to hire someone to help with the lift to hyper-v. Then I have a bigger budget, and I don’t have my nuts in Broadcom’s grip.

Reply

[-]

_BoNgRiPPeR_420@reddit

The budget ask is easy for a couple grand, but asking for a new FTE is a whole other ball game for us.

Reply

[-]

rune87@reddit (OP)

Unfortunately, I'm Govt and have limits on how much I can ask and morally, I honestly hate pissing away money . I recently just got another HelpDesk position approved so I'm not gonna push my limits if I don't need to. I was in consulting prior to this. I'm used to Hyper-V to VMWare lift and shifts and vice versa. Its just work.

Reply

[-]

jpStormcrow@reddit

I work for government as well. You need to accept prices are going up across the board and do does your board/CFO. I'm all for leaving VMware but these governments need to get off their fear of spending. That's why all these cyber security events occur in our sector daily.

Reply

[-]

rune87@reddit (OP)

Who said I have any fear of spending money? I do it, albeit judiciously, so I can allocate more to the cybersecurity side of the house. And it's been a highly successful approach. We pass all of our pentests and security audits every year with flying colors. There are more breaches in the Private Sector than the Public sector. Burning money doesn't equal security. If we want to loop VMware and Security into the discussion, we should all be taking a judicious long term look as to whether VMWare as a platform will maintain a decent level of security or if budget cuts and layoffs impact it as a viable platform or one riddled with Security holes.

Reply

[-]

jpStormcrow@reddit

11000 is pennies in IT software and you're thinking about switching to HyperV over it. My time and my teams time would be better spent working on security posturing or other projects than lifting two clusters to HyperV to save the equivalent of 10 business class laptops. I probably shouldnt have made the security comment as it was irrelevant to the discussion. I just have seen many branches of government in my ten years working for them fight tooth and nail to save a dollar then turn around and blow it frivolously on bullshit.

Reply

[-]

rune87@reddit (OP)

I was in consulting prior to this. I've either been working for or consulting to Govt for almost 18 years. I've worked with F500 companies. I've seen ungodly amounts of money pissed away because it was easier to someone than not changing....only for IT staff to get cut when bad times hit. And then more money pissed away for an alternative solution. I've seen and understand both sides of the equation. Yes..for us 11k is a rounding error in our overall budget. But I'm not looking short term. I'm looking 3-5 years out. Taxpayers bitch to high heaven with rate increases. Boards get questionable on spending. You spend in the good times so you can conserve in the bad times. If you develop a reputation for judicious spending, not just being cheap, when you go to them for the big ticket items, they tend not to blink. Its a balance. It's almost 45k a year over my 5 year hardware cycle. About 22k a budget cycle. What am I getting for that? What could I be spending that on? Extra redundancy? Training for the HelpDesk? Mini-Pentest? My ever growing Microsoft licensing bill?

Reply

[-]

jpStormcrow@reddit

You're free to run your department how you'd like to and I won't pretend to be in your shoes. I believe the fiscally appropriate thing to do is to perform the migration on your next hardware refresh and begin planning and training the team. This seems like a knee jerk temper tantrum over a rate increase - but so has this whole subreddit for the last month lol. I wouldn't even want to stress my team out over something like this mid cycle.

Reply

[-]

rune87@reddit (OP)

2024 Hardware refresh / Budget Cycle is part of what's forcing the debate now. I was expecting a much smaller rate increase. I've got 10 months to go before my licenses expire.

Reply

[-]

jpStormcrow@reddit

Probably should lead with that next time lol.

Reply

[-]

rune87@reddit (OP)

Yea...I went back and looked and realized I could have been a little clearer on the spec'ing hardware comment.

Reply

[-]

MajStealth@reddit

ill just be happy that we have a contract until 2025-07-15 we will see what we do then, maybe

Reply

[-]

_BoNgRiPPeR_420@reddit

Same, we just got our renewal quote in early November for our upcoming February renewal. $3800 or so for 14 cores licensed until February 2025. Gives us some time to evaluate our options.

Reply

[-]

Zander-@reddit

VMware vSphere 8 Essentials Kit (96 Cores included, no per core licensing. max. 16C per CPU or 32C per Host) can still be ordered and is still NOT listed as discontinued/EOL, even in the new January 2024 Partner-price-list. It's now a subscription model, yes - but the price is still very attractive with \~250€ per year.

Reply

[-]

dreadpiratewombat@reddit

I’m waiting for the day AWS, Google and Microsoft have the renewals for their various VMware services. There will be some fireworks I’m sure.

Reply

[-]

surloc_dalnor@reddit

Abandoning hardware for AWS and Kubernetes looks better every year.

Reply

[-]

wil169@reddit

That’s like the cost of 1 tower server give or take. Remember the days of 1to1 physical servers? It’s not that bad given the functionality and maturity.

Reply

[-]

wil169@reddit

Note I still think the deal sucks and shouldn’t have happened, but the alternatives aren’t great right now. Next year at renewal time it may be a greener field.

Reply

[-]

kiamori@reddit

Knew this was going to happen, based on what they did to LSI/3ware. We moved everything vmware over to Hyper-V and have not looked back. It's not that hard of a transition and everything is running smooth now.

Reply

[-]

Fog80@reddit

lol $5500? Our VMware bill is over 300k per year

Reply

[-]

rune87@reddit (OP)

I'm sure you are significantly larger than I am with a much larger budget as well. Would be interesting to see if the price increases have a linear scale to them.

Reply

[-]

Fog80@reddit

14 nodes/864 cores vsphere/aria/nsx

Reply

[-]

rune87@reddit (OP)

Just a little bit bigger. 😅 I'm really curious with everything licensed per core now, how that translates for y'alls licensing.

Reply

[-]

Fog80@reddit

We are already on per core subscription licensing. Bracing now for a 20-25% increase courtesy of Broadcom

Reply

[-]

rune87@reddit (OP)

Will be interesting to hear moving forward. I was prepared for a 100% increase. The 8x increase just was a serious wtf moment.

Reply

[-]

Fog80@reddit

We had another product acquired by Broadcom and saw about a 20% increase during the first renewal.

Reply

[-]

brewbrew@reddit

Scale Computing is where we've migrated customers to over the years from VMware, even before the Broadcom speculation. Support is spectacular, but it's hardware locked. They don't sell software-only at this point. It's essentially KVM with a nice frontend GUI.

Reply

[-]

fccu101@reddit

No real big discussions as of now - but we are evaluating going to Azure. We are %90 vmware as well - so it's going to suck. What hurts the most is that Broadcom is allegedly divesting Horizon. One of my major projects for next year is to start moving towards VDI.....I guess we could still stay on it or I can look at Azure Virtual Desktop.

Reply

[-]

cyr0nk0r@reddit

I'm a big fan of www.verge.io and anyone who says proxmox should really look at verge instead. That said, you're going from 700 per year to 5500. Honestly man, that's peanuts. Just pay it and move on. Invest your energy somewhere else. Keep this in you head the next time in your career you're in a position to replace VMware. Make sure their sales people know that you remember what they did. Vote with your money then. 700 to 5500 is not that time.

Reply

[-]

rune87@reddit (OP)

Thats per cluster. I have 2. Thats 11k a year x 5 years as my hardware cycle. Roughly a 45k increase. I can do a lot of improvements with that 45k. And that assumes VMWare maintains the same level of support and product quality. Broadcoms MO and past history says that's not the case.

Reply

[-]

cyr0nk0r@reddit

It's not your money. You're focusing more on the "principal" of the price increase. Costs for things go up. I think what Broadcom is doing is shitty too, but at 11k per year that's just not worth it to do anything about it. Now if you're looking to do a hardware refresh anyway or getting ready to replace your SAN and stuff like that, that is when I'd say have a hard look at verge or other HCI. But if you're embedded with VMware right now and it would take 250 man hours to change hypervisors, the economics of that just don't make sense at only 11k per year.

Reply

[-]

rune87@reddit (OP)

How are you coming up with 250hours of time? It's at the most 100 hours. I used to build environments for clients in under 30 hours. Even at $70 an hour, thats 7k. The business still benefits, and I get the benefits of getting a refresher. I don't stress about the work. I've got a great environment and don't run around like a crazy person so I have time to do stuff like this. Everything else in my environment is HyperV compatible.

Reply

[-]

cyr0nk0r@reddit

My man I'm not going to break down a hypervisor migration and how long it may take in your specific environment. I have no idea how many VMs you have but I guarantee you moving 2 clusters is more than 100 hours. Unless you have like 5 VMs. It's not just installing hyper-v and you're off. You've got to migrate each vm, test, have stakeholders test and certify. But whatever. If you're just here to complain about the price increases I'm happy to commiserate with you. It's terrible and I hope lots of customers move away from VMware because of this.

Reply

[-]

jpStormcrow@reddit

This guy gets it.

Reply

[-]

rune87@reddit (OP)

I always just respond because there tend to be a lot of blanket assumptions made in this group. We all have different environments, different experience levels, and just different requirements. We should be able to have technical debates. I like folks to be able to see that there is a range and to not be discouraged from tackling something just because a poster may have higher regulatory requirements than them. I've got a lot less red tape to work through than some places. Some enviroments I've worked on this would be a 3 year project. Its all a sliding scale.

Reply

[-]

micdawg12@reddit

Does not seem to be a community edition....

Reply

[-]

K3rat@reddit

We are 1/8th into hyper-v. We are working on preparing for Citrix infrastructure and MCS VDA migration. It is going to hurt but save us money in the end.

Reply

[-]

Chance_Brilliant_138@reddit

Yeah I’m 99% VMware shop and NOT looking forward to the new licensing bill. One of my tasks in the new year is to evaluate moving as much as I possibly can to Hyper-V. I’ll never be able to get rid of vSphere completely, but if I can get the licensing bill down to what it was previously I’ll be happy.

Reply