VMware flushes Windows vSphere client and Adobe Flash vAdmins will soon have just one graphical option with which to operate vCenter, after VMware decided to can its Windows client and replace it and other current tools with a single HTML 5 client. VMware's blog announcing the move points out that it's already adopted HTML 5 for several other products, but has held back on …
I wouldn't worry about that:
There are indications that there already was an attempt to make it possible to connect to individual hosts using a web browser with v6 -- call it web client for hosts, if you want -- but it was pulled, probably because it didn't work so well.
Dumping Flash cannot come soon enough, of course. But not having the safety net of the Windows client is... troubling. I think I'll give this a few months to mature before actually deploying it anywhere. Of course, that's what I would do anyway!
Couldn't agree more that Flash should die - no problems. Like there's what should be a reasonable WebUI now instead - great.
But - and maybe it's because I'm getting older - I seem to hark back to a golden age where there was full fat clients for managing systems. Where I can troubleshoot log files with ease, package applications up and not have to worry about browser issues.
And touching on the above comment - what about directly managing the hosts? Should vCentre be having one of it's bad days (says a VM is on, we all know it's not etc.) then how do you directly manage a host via GUI? Will this be embedded on all hosts as well as the vCentre server?
They already did a fling for an HTML 5 host interface for ESXi that is more functional than C# vSphere client direct to host ever was.
It already ships in ESXi 6.0U2 and most likely isn't leaving.
I think with its' success and after finding how rapid they can code with modern web setups instead of old legacy junk, they feel they can make the timetable countdown to VMworld work to ship a finished product (or at least announce and ship 60-90 days later which has been their typical timetable the last 2-3 times).
" I seem to hark back to a golden age where there was full fat clients for managing systems."
I agree, for the most part; not least because every single web interface transition I've ever seen has involved stripping out half the GUI features and forcing you to find them using obscure console commands.
Don't get me wrong - I love Exchange Powershell and all - but look at the difference between what you could do in an Exchange 2k7 GUI and the web-based EMC. Half the features are now gone, and if you want to do much more advanced that mount a DB or set up a new mailbox, you need to be going via PS now. Ultimately, Exchange flipping to EMC made me give up on using the GUI altogether, which is pretty much a sign that your GUI has failed to do it's job properly.
I had much the same experience with vSphere's web interface; I went from a full-fat client that worked to a shitty Flash front-end that mostly didn't, required installing a bug-ridden security nightmare (which inevitably needs to be kept 3-4 versions behind the latest because at least one of the horrible Flash apps only gets updated once every 3 years while Flash needs 8 patches a week) and forced me to re-write by security policy to allow a program that really ought to have been junked ten years ago. At least that's going away with HTML5. But none the less, I just continued using the full-fat client and putting up with it's buggy behaviour.
The thing is, I don't really see the advantage in a lot of cases. Oh, I get to avoid sitting through a 5-minute install once per PC, or going to the effort of RDPing onto server X via a VPN. Whoopie do. There's some use cases where this is a genuine improvement for the person using the software (plugging people into their email via browser rather than outlook, for example, is a real boon). But in many other cases, it just seems like going via a web interface either for the sake of it, or to make life easier on the developers (meaning half of them can be fired).
"Exchange flipping to EMC made me give up on using the GUI altogether"
This has always been Microsoft's goal. Making GUIs costs money, and Microsoft doesn't want to invest in this. It's much cheaper to simply tell the entire world that they will do everything through CLI and scripting and that they will like it.
What are they going to do, use a different vendor? Puh-lease.
You only need to care about ease of use when you're not the top dog. Now, over to VMware...
>> I just continued using the full-fat client and putting up with it's buggy behaviour.
Between the Windows and Web Clients, it's been my experience that it's the Web Client that's by far buggier. But it's also the one with all the snazzy new features -- everything since v5.5 inclusive, I believe -- so there's really no avoiding it completely, unfortunately, on a modern vSphere setup.
There has been a Host Client fling for some time which is the HTML5-based web client for ESXi hosts. As of 6.0 Update 2 this client has now been pulled into ESXi and is a fully supported feature (http://blogs.vmware.com/vsphere/2016/03/vmware-host-client-1-0-now-ga.html). The feedback on both the Host Client and HTML5 Web Client has actually been really positive so far but I invite you to make your own conclusions after trying either (or both!).
--Adam Eckerle, @eck79 [VMware employee]
The host client is acceptable, but the C# client still kicks its ass in a lot of ways.
As for the HTML5 vSphere client, that has a lot of growing up to do. If they can get it ready for even 6 months after VMworld, I'll be shocked. Even then, it's still not as fast or as responsive as the C# client. Right click and waiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiit.
Now, the HTML 5 client isn't as bad as that fucking miserable, piece-of-shit, eldritch horror flash fuckery, but then again, getting repeatedly tased in the balls is preferable to having to use manage more than a handful of systems with that thing. (And no, the 6.x release of the flashy fuckerdoodle didn't really make the boo boos all better.)
The HTML 5 client has promise. In principle, I far prefer it to a C# client. But there is a hell of a lot of work to do yet, and I am not remotely convinced it will ever actually be done.
After years of VMware shoving the flash client down our throats and telling us A) all our complaints are invalid and B) everyone likes it, really, it's just you that doesn't...VMware has lost all credibility regarding UI claims. The only company with a worse reputation in this regard is Microsoft.
So we'll see what's delivered. I, for one, have zero faith this will work out well for the actual practitioners.
vSphere is the hypervisor.
The "vSphere Client" is the software you use to connect/manage the hypervisor and vCenter.
For years we've used a windows application to manage the hosts and vCenter. VMware have now insisted that we use this shitty HTML5 abomination.
This article represents a collective sigh of disgust for vAdmins across the globe.
ESXI Free is a cut down version of the hypervisor and remains available, but you're probably need to manage it via the afore mentioned "shitty" web interface from now on.
"VMware have now insisted that we use this shitty HTML5 abomination"
To be fair, it is the flex client that was the abomination. The HTML 5 client actually has promise. If it is done on time. And if VMware listens to the concerns of systems administrators who aren't vapid sycophants.
The HTML 5 client is close enough that if significant engineer effort is put into finishing it, and solid QA is solicited from outside the "yes, sir" echo chamber it legitimately could be better than all clients which have gone before.
But will it be done on time? What is the matrix of which clients will connect to which versions of the software and manage what? How much of a complete nightmare is managing a multi-hypervisor environment going to be?
VMware is staying pretty mum on the details.
3rd party plugins (Everyone would rather write HTML5).
The C# client hasn't been properly updated for new features in years. Seriously if your using it today with 6.0 its time to stop.
The HTML5 embedded host client replaces the C# for fixing a broken vCenter. Its already GA (the newest build can be found on the fling site here).
https://labs.vmware.com/flings/esxi-embedded-host-client
The HTML5 web client for vCenter is on the fling site and can be found here.
https://labs.vmware.com/flings/vsphere-html5-web-client
The team is iterating quickly. The flash client will stick around briefly as it fleshes out, but its an agile team moving rapidly with quick releases. People are sick of flash (Security/Speed) and with the industry is preparing to deprecate flash (and HTTP) this is a serious customer demand.
"To have such a close partner abandon their Windows client altogether in preference for a browser based solution available on every platform has got to hurt."
I think MS don't really worry about a (relatively) little used client application for a rival to their own product.
Now that we have VMWare virtualizing OS's, what would you think about VMWare creating its own OS? That's mind blowing stuff! A bare metal hypervisor server virtualizing its own OS. Like a mirror mirroring an image in front of a mirror... That doesn't even make sense for VMWare... What would be the advantages of such OS? I guess that's what really VMWare is about, a small OS to virtualize other OS's, but how about a real OS, one that would be the most efficient OS in the world that would require other OS's to be virtualized...
VMware already has its own OS. It's called VMkernel, and it's what ESXi runs on. Hell, it's even binary compatible with Linux (and judging from current GPL lawsuit includes a chunk of code from it as well). And ESXi can supposedly virtualize itself, so nothing except the minor issues of usefulness and sanity stops you from running your services in a VMware OS inside a VMware VM inside a ...
("Bare metal hypervisor" is, by the way, a stupid marketing term that doesn't make sense on anything remotely resembling this kind of virtualization. No matter what you do, you are still gonna need memory management, scheduler, drivers, etc. - the makings of an OS. And in VMware's case you literally have a general-purpose OS with a standard kernel architecture etc. - just one that's optimized for running specific applications)
I'm no programmer, but how hard can it be to write a proper bit of software instead of some browser abomination?
or more usefully:
How hard can it be to write a proper web interface instead of some browser abomination?
There's a growing industry idiocy in attempts to replicate a rich application interface using JavaScript on a web browser. Whereapon it's no longer a web application it's invariably just a badly laid out and barely useable mess that just happens to be delivered through a web browser. Also, this isn't "HTML5", that's just marketing BS as HTML5 is nothing more than an extension of HTML4 and it's not a suddenly new or enabled environment to use with web applications.
http://www.theregister.co.uk/2012/08/30/vmware_new_interface_hands_on/
VMware has done a spectacular job of delivering the first enterprise-class Flash application I have had the pleasure to use. The web client in vSphere 5 was an interesting toy that we collectively poked at, shrugged, and went back to the familiar C# client. This time the training wheels are off, and VMware doesn't disappoint.
It bloody well does disappoint!
See, the problem is that the flash client works fine in small environments. It also works reasonably well if the server is some ridiculously overpowered Windows server running your vSphere server. If you have enough horsepower behind it, and you using only a handful of servers, it's not that bad. Even the first flash client.
But this all changes if you either A) use the virtual appliance or B) use more than about 8 servers. (Which, I hadn't had much of a chance to do when I wrote that.) When you place any sort of load on the vSphere server (or restrict its capabilities by putting it into a virtual appliance) it disintegrates.
VMware made it slightly less shit generation on generation, but there was something fundamentally broken about the flash client that meant the thing didn't scale worth a damn, and so nobody ended up using it.
So here we are with the HTML 5 client. 75% of the complaints I have about the flash client UI design go away either by A) making it look more like the C$ client, which was done in 6.0 or B) getting rid of Flash, which the HTML5 client does.
The remaining 25% is speed and responsiveness. The HTML 5 client still grinds to a fucking halt if you load the vSphere server up. The new 6.0 vSphere server already demands significantly more resources than the 5.x one. If they can't get the speed problem dealt with, I'm not looking forward to the "production ready" configuration that is going to be required to run the HTML5 based one.
And appliance is all you get now! So you can't solve this by throwing some dedicated overspecced windows box (or boxes, as you probably want a cluster of them) to run your vSphere server.
The idea behind the web client is great. Even the initial implementation at small scale was okay. Let's face it, the ability for the web client to queue commands so that you could keep working, instead of the "white screen of death" you get from the C$ client while it thinks is pretty handy.
But they just never ran the ball into the end zone. Now, they're telling us they're taking away the C# client altogether and sticking us with an unproven GUI with unknown bugs and unknown scaling and performance characteristics. Not cool.
I want the HTML 5 client to work. I want it to be what the Flex client promised to be. I want it to work with 128 nodes as good as with 8. I want it to be responsive and fluid and not make me feel an itch for the C# client.
One thing I have learned since way long back when VMware showed me that flex client for the first time: test everything in my own lab. Push the envelope. Never - ever - trust a vendor demo or a vendor lab environment. Push the tech to breaking. Do what the vendor says is unsupported.
So we'll see what this new one is like. Not with two nodes. Not in a VMware lab. But on the rack with dozens of nodes and everything I can throw at it. Maybe.
VMware has gotten quite touchy about people actually testing their product. So we'll see what we see.
I just assumed they had supplied you with hookers and blow ;) its the only way you could like that flash client.
I still remmber reading your original article and getting wound up.
The only reason they ever changed from the c client was because they changed from a company that developed software to one that aquired software and it was too much like hard work to keep c client inplace.
As you say T its about the buyers not the users.
Nah, I've never been "bought" by anyone. Some have tried - I've had honest offers of hookers, actually - but nobody is willing to meet my price.
If I sell out, I will lose all credibility and thus my livelihood. As such, in order to sell out, I need to be able to retire off of the selling out. In order to retire off of the amount made from selling out, I need to clear - after taxes - a little over $1.5M Canadian. This pays my debts, buys me a house on the island and gives me enough money to semi-retire and write my Sci-Fi novels.
In order to clear $1.5M after taxes I need $10M in gross income. I own 40% of my company and taxes will mean I only get to keep about 66% of what I make. Fiddling around with capital gains might up that some, but not appreciably, and shouldn't be counted on when factoring my "sell out" price anyways.
So, that makes my "price" $10M. If I am going to sell out, completely and thoroughly, to unreservedly say things I believe are not true and otherwise piss away my reputation and livelihood, I require a minimum of $10M.
To date, shockingly, no organization has offered me this amount of money. I know this because I still have to work for a living, I still live in this overcrowded shithole of a city, I still use Twitter, and my profile picture for every single site on the internet isn't a picture of my ass with the words "fuck you, world" tatooed in neon pink.
Not being bought doesn't mean I can't be snowed. And every now and again I just outright wrong. It happens. But totally in some vendor's pocket? That costs $10M.
Don't I ****ing wish.
Tell you what though - the HTML5 client is damn fast!
And much as I love the C#-client, I hate the white window when it gets it's knickers in a twist, The Flash thing is ok on a fast system but like anything these days with a flat, featureless, painful-on-the-eyes 'New, modern GUI for the modern age' you have to click more to do the same job, and the familiar has been buried in nested tabs.
SO I'm really hoping the new HTLM5 client can cut it. And if it's browser- and OS-agnostic so much the better.
If we have reached "Peak VMWare", it seems odd timing. A lot of admins I know still like traditional GUIs and hate the webclient with a passion (I never thought it was *that* bad). I guess I would have thought that given the rapidly diminishing gap in feature set and performance applicable to a lot of companies compared to Hyper-V, VMware would be trying to please the people using their products, not piss them off.
VMware doesn't care about the people using its products. VMware cares about the people buying its products. They are not the same people.
One group does as their told. The other make purchases after some hookers and blow. Anyone too small for hookers and blow isn't someone the sales team - or the company - cares about.
Top to bottom, VMware employees only care about the enterprise, and when you sell to the enterprise you don't have to care about the people who actually use your software.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
