Dennis no offence but you and Matt Dreyer sound like tech students. it like "oh we got a unusable tool. go to wiki.google.try-and-error.com to make your admin tool work"
What do you think is neccessary for succesful administration of a huge infrastructure.
You argue like this:
1. 10k users can manage 1 million hosts and applications from one single webpage
2. this single webpage should be able to scale and fit into cloud
3. every user should be able to high end customize his gui
4. and scale out should be achived with caching in the browser and in the inventory service
5. platformindependence, browserindependence and same enduser expirience on smartphone, desktop and tablets on different os
For me successful manage a huge infrastructure means:
1. a reliable client without refresh or most bad with login again because your performance boost in 6.0 seems to kill the ability that refresh is reliable too
I wokred with some guys in a lab on 6.0 and i made some storage change and i tried to refresh but it just did nothing show old data all time i had ro relog in and there we go refreshed data
Adittionally i have seen snapshot issues in 6.0 if acronis backup takes snapshot sometimes the vm is inresponcible after the api fail to create a snapshot
2. Standardisation, in a huge environment you need a standard ui everyone is familiar with and not something like "oh your button is in the left corner i didnt recognize it"
3. solid invest in education and slow changing of user interface
because our company and many others have spend lot of money in education and you started with 5.0 to bring out failed applications and gui to tell us next release "oh we did it in another way, please drop all your know how"
It is sad that the web client is unreliable and that we had to wait 5 years to get a par in speed and not equal in reliability or technologie state of the art client. (see chrome took away the NPAPI support and by the way chrome is the one browser which is not supported but was fast with this client)
But now you say "oh we will make it complete different over several years for companys with 1000 admins connecting simulatiously to 10 vcenters to manage 500000 user and 500000 vms"
To speak the truth i do not know such a company. and i know companys who have 3000 esx hosts there are not 1000 admins and they are not all managed in one single vcenter.
Because no one wants to deal with a single point of failure like this thats like your 4tb ram support in vms, deal with it thats a feature for max 1% of your customer same with your scale argument for your webclient.
I am doubt about the ability of the webclient to handle this workload because i cant use it in 2 man 2 host 4 vm lab environments
.
Dennis you and Matthew dryer should be aware of the fact you are the leader and you are not a cheap one.
If i want a student community project i go to XEN or Proxmox, if i payed thousand of dollars i want a usable product and not 2 years of excuses why the webclient is shit and why you choose flash.
We just dont care you have to deliver a useable product but you do not since 5.1 and i can tell you this means you dont get rid of your problems for more than one Year.
This tool is not Excel where anyone knows that it is enduser and shit.
Sorry for the clear word but we suffer for 1% of your customers and they have no benefit of this webclient stuff too.
So where is the sense
Never the less
Have a nice weekend and thank you for listening to us.
But you should make matthew listening and change his mind or maybe he want to go to oracle or microsoft gui team or best idea citrix.
ps please recreate a client that is independent of third party like browser or java updates.
Isolation makes a possibility to freeze the code fit for your environment.
And why you cant create a c# client that use loadbalancing in the db like
1k clients ------> loadbalancer (optional) --------------> one to several vpx on vcenter as much as you need to handle the 20k requests of the 1000000000k admins working simultaniosly -----------> sync replica db for scale out (one service on each db host which makes a refresh request on each active connection found in database for inventory changes push on demand not pull no autorefresh)
dont get me wrong but i am shure it is not possible to have a scaling client who needs two layers of cache and still achive real time data or get rid of refresh.
And last word i have noticed a bug that if you drop out vms of a vapp i cant manage them until i restart vpxa on the host. Is this a known bug? or a issue of my nested environment?
Kind regards
