@msff-amman-itofficer said in Open Source Hypervisors: do we really have them? do we really need them?:

my take on this:

ESXi free is limited, 8 vcpu per VM and that limit can be easily reached limit.

It's easily reached if you starve a VM of IOPS or RAM and it's spinning cycles waiting on IO. In reality VERY few things need 8 vCPU. I've seen The ONLY exchange server for 5000 users not need that many resources.

Given modern Skylake hardware, and 4Ghz Intel Xeon Cores's if your hitting the 8vCPU limit I'm REALLY curious why your not willing to spend the one time ~$200 per host that is the Essentials bundle to get some more features is a rounding error in your budget (it's like less than a $1 a day per host).

There's some more obscure things to consider.
On the VMware front...

SMP-FT (what witchcraft is this?)

How are heartbeats done, and is HA managed passively by staged scripts and the hosts, or a central server? (vSphere it's stand alone, works if vCenter down)

Storage HA heartbeats. What are they, and why do they rock? Why can't I use HCI storage for this?

What are fault isolation responses and why do you I need to know this?

Proactive HA. I have a server that is dying, why doesn't stuff move off of it pro-actively?!?

What is Virtual Machine HA, and what are API hooks for application HA?

A good book on HA and DRS by Duncan.
http://ha.yellow-bricks.com/

@texkonc said in Shots for Scott:

@wirestyle22 said in Shots for Scott:

@minion-queen I'm dying right now

with $150 worth of shots for Scott, yeah he may not fair too well.

Depends What and Where. My John Walker Black was $20 a shot last night. Then again This was where I was drinking it

@scottalanmiller said in Shots for Scott:

@grey said in Shots for Scott:

@scottalanmiller I didn't know there was a sequel.

• Bring It On (2000)
• Bring It On Again (2004)
• Bring It On: All or Nothing (2006)
• Bring It On: In It to Win It (2007)
• Bring It On: Fight to the Finish (2009)
• Bring It On: Worldwide #Cheersmack (2017)

It's an active series!

I knew about the first 2. My wife knew about 1-5 somehow (She was a cheerleader and Gymnast)

@scottalanmiller said in Hypervisor choice:

@Dashrender said in Hypervisor choice:

@John-Nicholson said in Hypervisor choice:

@jfath said in Hypervisor choice:
esxi is out because of the expense and their inability to provide a fully functional Web client to replace the thick client they've killed off.

Quick thing. ESXi has a free HTML5 interface (It's what replaced the thick client). When combined with VMware Remote Console (VMRC) for managing console sessions it's pretty damn poweful.

STILL, want a full thick GUI? Fusion/Workstation will do it and manage hosts.

Right - if the OP is willing to go to bat for a $700 XOA license, why not a $500/600 ESXi essentials license? or a $750 5Nines license?

Not comparable things. In this case, none of the licensing makes sense to me from any product. It's a charity that can use the money elsewhere and none of the paid stuff is needed across the board. But if you did need to pay for it, what you get with XO and what you with Essentials are totally different.

ahhh Non-profit. VMware is 40% off for this, Microsoft discounts hugely also (80%?). SCCM-VMM might not actually be that overpriced.

@Dashrender said in Hypervisor choice:

Something that confuses me - the company felt there was value in the 5Nine's paid product last year, but not this year - why not? If $750 really is making that much of a big deal to them, I have to wonder how solvent they really are.

Exactly. If this isn't within discretionary spending for IT, and requires some arduous budget approval something is really wrong.

@Dashrender said in Hypervisor choice:

@wirestyle22 said in Hypervisor choice:

@Dashrender said in Hypervisor choice:

@wirestyle22 said in Hypervisor choice:

@jfath said in Hypervisor choice:

@wirestyle22 Until 5nine decides to change the licensing model again. They completely lost my trust with the latest change - I assumed I would at least have the free version features if they increased the price yet again. They have moved full version pricing from <$100 to $350/core to $700 enterprise and killed the free version in the span of a few years.
If XO were to fork a closed source branch, we would at least have full source for the current version.

I spoke to someone claiming to be a VP and I told him that this decision is ultimately going to lose them customers. No one is going to pay what they are asking. I'd rather switching hypervisors

5Nines probably doesn't care. Those low end customers weren't paying the bills, so they dumped them.

How many future purchases will that kind of mindset cost them?

Again, probably not many that they care about. This is the same mindset that XOA is in. They don't want the low end SMB, because supporting those people costs them money, not making them money.

Outside of a 10% license cost increase in 5 years VMware's essentials (~$500) has been pretty cheap.

As far as everyone (5 nines, XO, etc) in this space having an intro price of $500-5000? This is pretty smart as it keeps you from wasting your sales people's time talking to people with no budget, and your SE's from supporting their POC's etc.

TO be blunt, if your working in IT and $500 is considered a material amount something is wrong.

1. It's going to cost you more in labor/time/opportunity cost to replatform.

2. $500 one time purchase spread over 3 years of usefulness is 45 cents a day. 6K over 3 years is $5.40 per day. My wife spends more on Starbucks, and Scott and I average far more than that on Scotch. Assuming your IT labor has a full carry cost of 100K (Cost of benefits, cost of office space, cost of equipment). If paying for some decent management and monitoring and alerting tools save you 5 minutes a day (beyond reducing and helping remediate outages) then even paying 6K is well worth it.

3. Lastly, if this REALLY is a huge amount of money for a business that you are a FTE I'd encourage you to stop what you are doing and start applying for a new job. Either your company isn't healthy to the point that your job is at risk, or you are at risk professionally by working somewhere that would have IT people waste huge amounts of time. By the time they fail or you end up wanting to leave it may be too late. As a hiring manager, I interviewed a lot of SMB burn out's who's skills were on 10 year old operating systems, and their management process's lacked any scalability because they had never had the right tools.

Note I'm not saying don't go deploy or use free software. I'm just pointing out that is your willing to radically change your infastructure because you can't get $500 in budget that's a clear warning sign about your buiness.

@jfath said in Hypervisor choice:
esxi is out because of the expense and their inability to provide a fully functional Web client to replace the thick client they've killed off.

Quick thing. ESXi has a free HTML5 interface (It's what replaced the thick client). When combined with VMware Remote Console (VMRC) for managing console sessions it's pretty damn poweful.

STILL, want a full thick GUI? Fusion/Workstation will do it and manage hosts.

@Tim_G said in Managing Hyper-V:

@bigbear said in Managing Hyper-V:

@JaredBusch in the case of the domain being down can you still log in locally?

Same as any Windows server. There's domain logon and local user logon. Also, as matteo said, cached credentials.

Not to mention "other" ways if you have physical access to the server, or remote with iDrac/ilo.

Ransomware. I've seen cryto attack that encrypted all the VMs

@Tim_G said in Managing Hyper-V:

Turns out, WebVirtMgr was too good to be true. I couldn't get it working on Fedora 26 or Fedora 25. Hours wasted.

I looked at Proxmox, but that's a Debian "appliance". I'm not using Debian in enterprise and don't want to. No time wasted, didn't bother.

oVirt wouldn't even install on Fedora 26 or 25. Apparently it's built for Fedora 24, I'm not going there. Even then, it doesn't seem like it would install. Time wasted trying to get it working. Packages were updated as of yesterday, so I was thinking they would work. I was wrong.

Briefly looked into Scale... I don't see any package or rpm to install. They want you to use an appliance or something. Not interested.

So much non-working stuff. There were others, but not for enterprise Linux such as RHEL/Fedora based. Everyone tries out their projects, and just abandons them, or they make it so hard to install and it's just insanely unstable.

I'd rather it be like the Windows ecosystem... where it either works great and has great support, even free versions, (StarWind for example), or it simply doesn't exist.

Moving on, I did find Kimchi. I don't remember if that was mentioned in this thread, if so, thank you... but I found it mentioned somewhere on the net.

Kimchi... easy to install, webpage up and working out-of-box if you know what I mean. I don't know if it "technically works" because I'm testing it on Fedora 25 that is a VM itself... but at least I am able to log in to the web page, and see a nice what-seems-to-be-working interface:

Well... it was working. I just went to grab a screenshot to show how great it was (what was I thinking?).. but now this:

I may be able to get it working better on a fresh install, so I won't dismiss it yet... but still, no viable options on Linux for web-based VM management / console access either.

It may be financially better to install Hyper-V (free) and pay for 5nine... labor costs a lot to mess with all this non-working and non-applicable stuff.

At $600 one time fee for vSphere essentials starts to sound real nice about now also. He'll even SCCM-VMM at ~8K isn't awful compared to something that MIGHT work on your OS today but not survive an upgrade.

It's kinda bizarre because enterprises with large scale pay for managemnt tools (and are better positioned to write their own/script stuff). SMBs where they lack the in house bench and depth to operalize free are the ones always trying it. Now to be fair enterprises getting 80% off because of their scale, and they are better poised to leverage some features do shift the value to running comedical software.

Im curious why you don't like appliances?

As a consultant when I saw a solo SMB admin wasting 60% of his time messing with stuff that should have been taking up 2% it was hard to Disagree with the management decision to replace him with a MSP.

@scottalanmiller said in Managing Hyper-V:

What's funny is, so often they just drive people away. They've made me, without owning an XBOX for years, truly hate the XBox and all MS gaming, uninstall Skype from everywhere and totally ignore S4B (do they still offer that, Teams replaced it!)

Teams didn't replace S4B. Teams is a Slack Competitor. S4B never had a functioning multi-user chat.

@scottalanmiller

Because Microsoft has a long history of undercutting their ecosystem by...

1. Breaking API's on purpose to sabotage people (Novel admins still hold a grudge).
2. Competing with partners.
3. Locking out partner ecosystems (Azure/Azure Stack freezes out backup providers that don't run in guest VM agents).
4. Screwing with partners and customers by changing PnP to crazy pants stuff (See VDI licensing that caused MVP's to quit the program, Azure Stack pricing that will FORCE opex pricing even for private cloud deployments with your own hardware). It's like they watched VMware with vRAM and said "hold my beer".
5. Killing products that support the SMB space in ham-fisted ways.
6. Their slow attempt to break Skype and force me to use S4B that I'm still annoyed with
7. Creating products that need 3rd party ecosystems with unique requirements for hardware/software (Storage Spaces) then abandoning that product (Storage Spaces Direct abandonment of Shared SAS backplane).
8. Complete lack of quality control on the ecosystem. a 5 pack of American Light beer and a back rub will get your drivers signed.

@Tim_G I've seen something that looks strikingly similar for ESXi (Maybe it's just every angular node.js app looks the same). Note these UI's are REALLY limited to only the most absolute basics of management (Partly because it takes YEARS to build a "full feature" UI for the 200K+ API calls that a hypervisor platform can perform, and when done on a platform with notoriously changing API's (KVM historically) this only works when you control the full stack (and even then Scale and AVH both are still focused on the 10% that are 90% of your workflow).

@scottalanmiller said in When to use VMWare over free hypervisors?:

@matteo-nunziati said in When to use VMWare over free hypervisors?:

@stacksofplates like let the san use uts own snapshot/backup sw?

Yes, which isn't quite fully offloaded, but a reliable means of communications so that the SAN can do the majority of the work.

Actually vVols is different than SAN Snap-ins in how snapshot offload is done. VMFS snapshots do not exist at all in the case you use them. You also get a management, and command integration so you can set compliance requirements (encryption, retention, replication, performance minimums) and apply policies automatically (using OpenStack, VCD, vRA etc) and it will validate that those configurations are continuously met and alert if not. Nifty stuff.

@Reid-Cooper said in When to use VMWare over free hypervisors?:

You would not likely buy something just for this, but VMware has vVols. I don't think that anyone else has an equivalent technology.

Vvols is amazingly powerful, and there's nothing even on roadmap for something similar from others. Even small shops can benefit from the 50x snapshot performance improvement.

@matteo-nunziati said in When to use VMWare over free hypervisors?:

@Francesco-Provino said in When to use VMWare over free hypervisors?:

@matteo-nunziati said in When to use VMWare over free hypervisors?:

@scottalanmiller said in When to use VMWare over free hypervisors?:

@matteo-nunziati said in When to use VMWare over free hypervisors?:

@scottalanmiller said in When to use VMWare over free hypervisors?:

@matteo-nunziati said in When to use VMWare over free hypervisors?:

@DustinB3403 said in When to use VMWare over free hypervisors?:

The cost of the solution isn't expensive if your business requires those features.

example?

Fault Tolerance with vendor support for it. Technically not limited to Vmware, but essentially limited to it. I believe Suse with Xen is the only other vendor who offers OEM vendor support for that.

Agree 100%. It is one of the cheapest supported solutions. Issue is if you can afford it! Usually not here.
Of course exceptions can be around. But are exceptions imho in the small business.

Not sure it is the cheapest. Compare to Red Hat, I bet RH is cheaper. I've not compared, I'm just guessing.

Not really. If you want live migration and other stuff you have yo go rhve or how the hell they name ovirt.

Plain KVM has done live migration from the very beginning…

Yes but if we talk about support for this kind of things, this is provided via rh virtualization.

Try live migrate with stock centos: it doesn't work. You need to pull from rh virtualization sources. Centos has a set of packages for this specific need. It also does share nothing migration

Also long distance vmotion is another fun on. On vSphere I can migrate stuff from London to NYC without disrtuption. Also NSX will virtualize the network so that VMs network will actually work running active active over long distances with full ingress/egress and BGP.

@scottalanmiller said in open source hypervisors: do we really have them? do we really need them?:

@matteo-nunziati said in open source hypervisors: do we really have them? do we really need them?:

KVM/libvirt is basically a Red Hat show. If Red Hat will drop KVM there will really be someone which will step up and will continue the development?

It's not owned by or controlled by RH. RH is not likely to drop it, less likely that MS dropping Hyper-V. Knowing that someone else will pick it up and that all they will do is lose control is one of the many benefits of open source to us, the consumers. It keeps RH from dropping things in a way that we don't have protection with for closed source.

KVM is part of Linux, not RH. It's heavily contributed to by Canonical and Suse but, more importantly, IBM. Even if RH walked away today, KVM is not in the slightest danger. If MS did that to Hyper-V, it would be over - period.

So yes, the open source nature here provides us the most extreme level of benefits and protection that exist in the industry.

Who outside of google project zero is really doing this though? No F1000 I worked for ever did their own open source audit.

@scottalanmiller said in open source hypervisors: do we really have them? do we really need them?:

@FATeknollogee said in open source hypervisors: do we really have them? do we really need them?:

Just want to make sure I'm following this correctly. Is this the "Xen" that you guys are referring to ? https://www.xenproject.org/

If yes, what "GUI's" are available to manage Xen?

XenCenter, Xen Orchestra, OpenStack, AWS, RS, and many more. No shortage of options

OpenStack is really a CMP, than just a hypervisor GUI.

@scottalanmiller said in When to use VMWare over free hypervisors?:

I get that they were trying to build hardware FT to try and compete with VMware FT, but...they failed big time.

Except that kind of hardware is already dead common and has been for decades from every major vendor AND they were not aware that Vmware (and Xen) already had this!!

To be fair to FT, it doesn't do RDMA transport (yet) for the memory, and it has a scaling wall at 4 Cores. Some of these weird hardware solution can scale a little bit better.

@scottalanmiller said in When to use VMWare over free hypervisors?:

@scottalanmiller Redhat Costs more than that just to support a single server. (Seriously, get a quote). Even SuSE isn't that cheap. $1200 for 3 x 2 socket servers 24/7 support is wildly cheap for a hypervisor (Note I will give credit to Redhat and SuSE is they will also provide support for Linux as an OS for VM's with their higher support bundles so that will get you some OS support which is damn nice, but even then more people run RedHat and SuSE on ESXi than KVM rather than pay the premium to the linux vendors.

Right, in that situation, you assume paying for support for the OS already. So those costs are already additive to VMware's costs in most cases.

It does require a higher level of support to get the hypervisor stuff (at least used to with RedHat) vs just basic instance based support. I think there's a Windows Datacenter like datacenter cross point they might just throw it in.

Again, most people who are paying for Redhat's premium licensing and running it at scale likely care very little about the cost of management/monitoring tools vs. solution stability/density/uptime etc.