@scottalanmiller said in When to use VMWare over free hypervisors?:
In reality, AD and Windows are overly obtuse for most SMBs. But they are accessible and sold heavily so people claim knowledge and play with them.
Hmmmm For a while there I thought I was the Golden God of GPO (A skill I'm happy to have not used for 2 years now!).
@scottalanmiller said in When to use VMWare over free hypervisors?:
It's more that there is a small space in the SMB where it is beneficial, but a HUGE space in the non-SMB where it is. If I'm running a Fortune 1000, I'd almost always choose VMware because the cost is small compared to the workloads, the support is assumed as a cost no matter what and they make the best stuff.
We have a slide deck somewhere that shows the cost of an Oracle RAC deployment and has a bar graph of all the costs. The VMware licensing piece is so skinny it looks like a rounding error. There are a lot of area's in IT where for a 1-2% cost you can get some damn nice management functionality. My favorite example is people forgetting to pay for iDRAC/iLO. I've argued with the product managers they should have a special SKU/Model number that it is IMPOSSIBLE to remove from the quote (it's just rolled into the motherboard) so that procurement clowns can't strip it from the quote to "Save money!"
@scottalanmiller said in When to use VMWare over free hypervisors?:
Is that really true? So many people use it without understanding it. The "using it because it seems easy" mindset makes for a support nightmare, similar to what Windows faces with their ecosystem. So many things are done incorrectly because it seems like you need to knowledge to do it. Then everything blows up.
Obtuse design to keep amateurs away from a product work if your willing to pay the non-trivial opex costs (which are huge). The only reason this works well in a SMB is if you want it to be painful for anyone to drop in and replace you or outsource you.
When I drop into a SMB with 2 servers that are running NetBSD and Gentoo (The ricer of linux) and are not using QWERTY on the console I know someone tried to make himself impossible to replace...
@scottalanmiller said in When to use VMWare over free hypervisors?:
@Tim_G said in When to use VMWare over free hypervisors?:
@John-Nicholson said in When to use VMWare over free hypervisors?:
@Tim_G I'll take it. My dog requires 3 walks a day, and play time. My cats I had could be ignored for a week or more without much effort given enough food/water and fresh litter was left out.
True, but they don't do anything useful without an insane amount of training, time, and money ^_^
And the last time (admittedly some time ago) that we hired VMware training, the class trained the VMware staffer because Vmware didn't know its own product. I've had a certain lack of faith in it ever since Xen and Zones folks were the ones teaching VMware how to use its own software to the "expert" that VMware had on staff.
I'm sure Vmware has loads of great people, but even Vmware stuggled to find what I'd call competent users internally whereas finding people who knew Xen was pretty easy (and still is.)
I suspect this was in the early days and you got a bad teacher. I've had pretty decent trainers for the classes I've done. I'm also helping with some of the curriculum right now. For a product (vSAN) that releases two to 3 times a year it's "fun" keeping this stuff up to date.
As far as training new employee's on the product that's a reality of any company with 20K employees. You will hire people who don't know something and train them. I've seen the NSX team hire people with no networking background but deep Virtualization/VDI just to offset the amount of hardcore networking people who lacked exposure outside of networking (it's a weird product that has to startle a LOT of vertices). Especially for SE's you tend to hire people and give them a training plan (you'll take these classes, do this self study, know these skills at this level by yyy). When you work for a 20 man firm you can hire purple squeals. When your trying to staff an 8000 man field sales force, you have to accept that you need internal training programs.
@scottalanmiller said in When to use VMWare over free hypervisors?:
@John-Nicholson said in When to use VMWare over free hypervisors?:
@scottalanmiller By this logic you should always use DB2 over Microsoft SQL, zOS over Linux or Windows, ARM or Power Processors over X86 and Juniper over... Well anything that isn't so damn weird and complicated as JuneOS.
IT naturally gravitates to commodity platforms for general purpose non-speciality stuff.
What was the statement to which that was a response? LOL
Windows Support issue.
@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.
@scottalanmiller By this logic you should always use DB2 over Microsoft SQL, zOS over Linux or Windows, ARM or Power Processors over X86 and Juniper over... Well anything that isn't so damn weird and complicated as JuneOS.
IT naturally gravitates to commodity platforms for general purpose non-speciality stuff.
@Jimmy9008 said in When to use VMWare over free hypervisors?:
@John-Nicholson said in When to use VMWare over free hypervisors?:
@Jimmy9008 Application clusters require exponential more complexity in support, monitoring, and operations (patching, troubleshooting). I've seen SQL clusters cause more outages than they solved in many cases. Hypervisor HA is VERY simple in comparison. App HA in some cases (AAG, Oracle RAC) has a VERY steep entry price (quickly gets over 100K). If my operations teams are not trained/certified/skilled on these solutions I could be extending outages, or extending costs for basic tasks.
An aircraft carrier is a superior solution to a Catamaran except when you only have a 4 man crew who never were in the Navy...
While I love App HA, and many people need it. For some Hypervisor HA is a good middle ground.
I somewhat agree. It depends, as always. The example I was giving is not complex and should be a starting point if you are thinking of HA for IIS and SQL Server. Just jumping to hardware level HA using nodes with VMWare or nodes with HyperV or whatever is just lazy. Think about application level HA too as a start point.
It's not lazy it's considering licensing of application stuff, as well as operational costs. Back in the day Hypervisor HA was considered exotic and expensive (and it often was). Now it's mundane (tons of ops people know how to deploy/support it), and cheap (Compared to application HA clustering on licensing and opex). Now SOME app's are cheap (DHCP/AD being examples), so there is some thought I agree but it's not lazy for someone with a mixture of apps and things who wants to be able to do basic hardware lifecycle management without disruption and hypervisor patching (which is painful otherwise on a monthly basis) to go straight for Hypervisor HA then evaluate what next?
Past a dozen VM's rebooting EVERYTHING just to do a host patch gets REALLY annoying to fully validate everything.
@Tim_G I'll take it. My dog requires 3 walks a day, and play time. My cats I had could be ignored for a week or more without much effort given enough food/water and fresh litter was left out.
@matteo-nunziati SMB's have far more diverse needs than you would think.
I worked for a 50 man call center where a missed phone call could mean someone did get a lung transplant (We did dispatching for Organ transplant). Don't assume everyone in SMB is a small retail shop who could loose their computer for 4 hours and loose $20. I'd argue larger shops often have greater tolerances for downtime than small shops in some cases because they have operational contingencies.
That said, if $1200 a year for flat rate support on 3 hosts that can easily run 100VM's is too much I'd question why the entire project of what your doing is viable. You can't get support at that pricing on any other platform I've seen.
@Jimmy9008 Application clusters require exponential more complexity in support, monitoring, and operations (patching, troubleshooting). I've seen SQL clusters cause more outages than they solved in many cases. Hypervisor HA is VERY simple in comparison. App HA in some cases (AAG, Oracle RAC) has a VERY steep entry price (quickly gets over 100K). If my operations teams are not trained/certified/skilled on these solutions I could be extending outages, or extending costs for basic tasks.
An aircraft carrier is a superior solution to a Catamaran except when you only have a 4 man crew who never were in the Navy...
While I love App HA, and many people need it. For some Hypervisor HA is a good middle ground.
@Jimmy9008 This is list price FWIW.
Also Production support (24/7) is a VERY low cost to add on and worth it for 99% of people. I know some people may not value support, but when you've got an issue at 3AM forums and Slack are not as helpful as you'd think.
I see them a lot more often. Oil gas is GREAT about building applications that if they fail people could die or millions are lost, but they lack native HA.
I saw a OEM build a vSAN cluster with FTT=2 RAID 1 (3 copies of data) and THEN ran FT (So 2 copies of VM's with a shadow VM) so 6 copies of data! I asked them why and they mumbled something about people could die so I dropped the issue of wasted capacity 
@Jimmy9008 said in When to use VMWare over free hypervisors?:
Where did I say 100% uptime? I didn't. 12 VMs is small. Everywhere has downtime, yes, but this isn't complex and for IIS and SQL Server etc... this is not an unreasonable setup. Not hard to manage or design either. I'm shocked y'all think its suck a crazy setup.
Why I like having hypervisor clusters even for app clusters is it let me do host maintenance in the middle of the day when resources are cheap vs at night where they are expensive (overtime pay, comp time, or just burning out my operations staff). Also Proactive HA will detect a host is failing (FANs fail, thermal warnings, memory errors) and evacuate a host and put it into quarantine. Still collecting stats on what % this prevents on host failure but you'd be surprised how many hosts give off warning signs that they are failing.
Full disclosure I work at VMware blah blah blah words are my own...
Microsegmentation with selective layer 7 service offload (NSX).
Platform residency like Pro-active HA (as well as vSphere HA and VMFS being far more resilient and tunable than other platform's options).
Platform features (vSAN isn't the only local storage replication system, but it's more tightly integrated and offers more data services and maturity than most systems out there).
Granular performance controls (DRS's algorithms are unmatched in the industry, NIOCv2, SIOCv2, highly mature APD/PDL detection at the storage PSP layer that would require a 3rd party software on most platforms). Using all this stuff I typically see 40% denser host utilization than people using purely free solutions (or just ESXi free).
Wholistic monitoring solutions and integrations (vROPS, Hyperic and LogInsight have crazy amounts of out of the box functionality across everything from hosts, to applications, to networking and external devices).
Hybrid cloud. Your replicating to a VMware partner in the vCloud Air Network like OVH or Softlayer (VCAN), or wanting to run Hybrid cloud operations to AWS (coming soon (TM).
Operational reasons. I can throw a rock and hit someone who knows how to manage ESXi and vSphere. There are a bazillion people are trained and know how to do not just basic Install configure manage, but also advanced troubleshooting.
Low cost 24/7 Enterprise support. I can support 3 hosts with 24/7 phone support for ~$1200 on an essentials plus bundle. Microsoft's lowest flat fee support option I've seen is 40K a year as part of an ELA.
Lost cost (Essentials Plus is ~6K. For similar functionality I'd need to buy SCCM VMM which costs more and lacks a 24/7 support option).
Better driver/firmware quality control. VMware VCG doesn't blindly certify anything for a 5 pack of heineken (Realtek was banned for a VERY long time). Do to what customers use the drivers for (Never consumer use cases) they get more attention and there is a higher expectation.
The vDS is incredibly advanced and mature. Beyond NIOCv2 shaping functionalities that are a check box away, Advanced LACP, single click to setup CDP/LLDP send/receive. VERY rich load balancing algorithms hashes (not just a basic IP hash). Other platforms historically require 3rd party NIC vendor dependent tools do things like this.
Security/Compliance. There are platform features that are painful to replicate in other places, or if doable require endless amounts of scripts or 3rd party projects that may or may not be stable. There is a full DISA STIG for classified use case. Due to ESXi's tiny size (few hundred MB for the kernel) it's patch surface is tiny. Compared to platforms that minimum install is 20+GB who require monthly patching to remain compliant.
Mature native backup API's (VADP/CBT), Native Write Splitting API's (VAIO allows near zero RPO) and all kinds of fun platform features that allow for a rich 3rd party ecosystem.
Rich API's, and SDK's that are stable and managed by grown up's (This isn't docker where everything breaks every 3 months).
VMFork plus Photon allows for linked clones VM's that have zero net Memory/CPU/Disk overhead and can be booted in 400ms. This can be controlled by Kuberentes or Docker endpoints and gives you the benefits of containers, with the isolation of VM's (Performance control, microseg, visibility) with vSphere integrated containers.
Free OpenStack deployment that's a simple wizard to setup (vSphere Integrated OpenStack).
Also, ESXi is free and offers a nice HTML5 UI that's quite easy to use.
Anecdotally, it's just easier to deploy and manage, and tends to do weird things a lot less (and when it does you've got support to reach out to). The new VCSA vSAN bootstrap system, and the configuration assist for bulk setup is awesome. With a simple wizard I can have my entire cluster deployed in under an hour.
@scottalanmiller I think people who are payed low wages sometimes internalize that that is all they are worth (and SMB managers often try to instill this "don't view yourself as valuable or increase your value because I can't pay you more!").
If your feeling this go listen to Katy Perry or something...
@scottalanmiller said in Negotiating for a Job You Are Leaving:
@wirestyle22 said in Negotiating for a Job You Are Leaving:
@scottalanmiller said in Negotiating for a Job You Are Leaving:
@matteo-nunziati said in Negotiating for a Job You Are Leaving:
@scottalanmiller said in Negotiating for a Job You Are Leaving:
@EddieJennings said in Negotiating for a Job You Are Leaving:
If the employee has zero desire to continue their employment at the company, would it not be a waste of everyone's time for the employee renegotiate the position? This is assuming there is no offering that would rekindle said desire.
If you truly believe that there aren't enough monies and benefits in the universe to convince you otherwise, no don't waste the time. But that's never really the situation. This is a job that you were okay with at a current salary today, but a change to that job tomorrow would make no salary good enough? While theoretically possible, it's not realistically plausible. This may happen once or twice in the whole of human history.
add one: I'm going to leave next year after a renegotiation last year. and for sure no one will pay me more. also, it is not sure I will quict with a new job agreement already in place.
anyway the main reason I've stayed another year here was not more money (even if they offer me and I accepted) but more time flexibility. I think that if money is "enough" better time is always the thing to attain.
That's why I was careful to add benefits, I agree. I've taken a 90% cut in pay over my last corporate offer to have a lot more time with the family and freedom to do what I enjoy.
Not everyone has the luxury of making that choice though. You have a very unique situation
Most everyone has the luxury, very few take it. Had you asked me before I did it if I had that luxury, I'm sure I would have said "no" too.
There are so many poorly qualified individuals at the upper levels of IT in large companies I'm pretty sure a mildly intelligent coconut would have the opportunity at the right time. Anyone with a room temperature IQ or higher can make 6 figures in IT somewhere.
@scottalanmiller Verizon operates this way (Although Telecom and competence are kinda a mess) With certain compensation levels being done based on if your a band 4 (you go down in band as you get closer to the CEO).
@scottalanmiller Always loved CS Lewis's quote on this.
It is a mistake to think that Christians ought all to be teetotallers; Mohammedanism, not Christianity, is the teetotal religion
