So can someone at vmware possible confirm that KB2004784 has wasted a bunch of my time and actually no longer applies to ESXi 6?
I just want to boot a ESXi server off of USB and not rely on a ramdrive. I grew my USB key to 16gb, yes, gigabytes. Considering only a few hundred MEG are being used - hey I think 15 gigabytes is darn enough for SYSLOGS!!! Why did all this stuff with vmware get so unnecessarily complex, was it just to support their certifications? I've been using vmware since workstation v2 and cannot believe I have to get out a calculator and run partition commands for something so simple but even then why does this KB not work?
So after hours what should be a 2 second thing, before I just pay the $50 for a real hard drive can a vmguru actually help with the VERSION 6 way to fix it? KB2004784 seemed like a perfect solution - edit one config and reboot and let the boot process auto-partition my 16gb usb and allow it now persistent for logs. No? Not with v6? What's the new supported method on v6 to use a 16gb key to log to the key no not a shared storage vmfs datastore the 15+gig left on the boot key and no longer use an absolutely unneeded RAMdrive?
Any help greatly appreciated.
BTW, I have even re-build using formatwithmbr and this is my layout after making the change in the KB above and rebooting as you can see, a 16gb key and no scratch partition:
Thanks I already found that however it doesn't answer the question or address the issue of a broken installer in v6 and a misleading KB article that I am waiting for an official confirmation actually does not apply (by design) to v6.
I am not looking for that manual workaround per say, and even then that article about manually creating the scratch is poorly written and have you read it start to finish? If you print it, it's a LOT of pages for a supposed "simple" task. I'm pushing my 30 yrs in the business and at some point, the vendor either has to correct a broken process or write a step-by-step workaround. I literally wasted days of my life on the last vSphere 6 patch that quietly broke iSCSI communication along with the CBT API issue.
It's funny that they keep doing this - fixing it in v5 then taking out working code. If they really did take out the above KB solution in v6 that was a logical one step fix in v5, perhaps someone at vmware can answer WHY on earth they would do that. Also, their KB all over states anything over a 8gb usb automatically gets a scratch partition - so either they need to fix all the inaccurate KB articles to exclude v6, or patch to match the docs, right?
Seems odd that Unix is like 40 yrs old and with all the complaints I've heard over the years about Windows, it would take that much work to auto-partition correctly based on whether the usb stick is > 8gb, LOL. In fact, when I look up how to verify the partition it's surprising how many Linux sites literally after saying run the commands, get out the calculator to get the sectors, mult by 512, etc. then they literally say "or just plug it into windows and open disk manager" since outside of the FORTY yr old Unix, MS figured out how to write code to display your simple partition table in readable numbers! 
Again, thanks for the info but I need a better explanation on why it doesn't work properly the way the docs state it should.