Windows 10 1909 EOL

[Archon Prime]

Sigh.... so I saw this on my feed this morning https://www.bleepingcomputer.com/ne...ws-10-1909-reaches-end-of-service-next-month/

I have 2 machines still running 1909 and the damn things will not update no matter what I've done, shy of reinstalling Windows 10 with the latest v2009 from scratch. Windows update has come up several times showing that the 20H2 update is ready for this unit. Attempting to install goes through but on reboot, nothing is installed. Tried installing 2004 and 2009 from the media creation tool, Windows 10 updater, etc. It shows it installing forever (a good hour +), then after the reboot is requested, it's right back to 1909 with the error code in the update history: 0x1900101. Looking up that error and trying everything to no resolve. All other updates pushed to the 1909 Windows 10 are installed no problem.

The one machine I've been trying to upgrade has had all drivers and BIOS updated to the latest, all fixes, purges, sfc, DISM, etc done to it that I'm blue in the face. I'd very much prefer not to be going a reinstall on these if I absolutely have to. I'm just really not sure if anyone else has any steps they might suggest in regards to getting these up to date.

The main one that I've had issues getting past 1909 is this one:

HP Slimline Desktop PC 270-a0xx
Running Build 18363 (64-bit) 1909
AMD A9-9430, Radeon R5, 3.20Ghz
8GB RAM

I've updated all drivers and BIOS on this machine directly with HP Support Assistant for that machine and as well checked over it with SDI origin to ensure nothing else was missed.
Peripherals connected direct is a Brother Printer (USB) and a Flash drive for a local backup alongside their cloud backup.

 

[Diggs]

I've had a few straggler Win7 machines fail to update to 20H2 but had no problem upgrading to Win10 via 1909 and then to 20H2 through Windows Update. Makes me go hmmm...... Windows Updates are so much better than the Win7 days but still seem to have quite a few bumps in the road.

 

[Archon Prime]

I've had a few straggler Win7 machines fail to update to 20H2 but had no problem upgrading to Win10 via 1909 and then to 20H2 through Windows Update. Makes me go hmmm...... Windows Updates are so much better than the Win7 days but still seem to have quite a few bumps in the road.

Yeah no kidding. I have 30 endpoints I manage. Only this desktop and 1 laptop are not installing the latest. The desktop just makes me scratch my head and from what I know of, I've done what I believe is everything I can think of. I haven't actually looked at the laptop since it's usually in production when I have a minute to look at it.

 

[Diggs]

Yeah no kidding. I have 30 endpoints I manage. Only this desktop and 1 laptop are not installing the latest. The desktop just makes me scratch my head and from what I know of, I've done what I believe is everything I can think of. I haven't actually looked at the laptop since it's usually in production when I have a minute to look at it.

There is some hardware that MS blocks from upgrades until it has the problems worked out. I used to keep track of the details but not so much any more.

 

[nlinecomputers]

There is some hardware that MS blocks from upgrades until it has the problems worked out. I used to keep track of the details but not so much any more.

If I understand the OP these systems are being offered the upgrade so they shouldn’t be on block lists. They’re just puking on the upgrade and rolling back. I‘d just nuke and pave them. Most of the time it’s faster and you get a more stable system in the end.

 

[britechguy]

I‘d just nuke and pave them. Most of the time it’s faster and you get a more stable system in the end.

And I agree, given the background that's been given. I never do N&Ps as the first choice, but if you've already taken all of the usual intermediate steps that would make an N&P unnecessary, and all have failed, then tearing off the proverbial band-aid and just doing an N&P makes perfect sense.

And for a system that's been this recalcitrant about updating, it's well worth starting with a clean slate. You'll probably never have the same issue again.

 

[Sky-Knight]

Remove the peripherals and try again, I've had some systems do stupid things on feature updates with USB storage attached.

 

[wavey]

I had a 1909 that would not upgrade just last week, same as you it looked ok until reboot, when it reverted back to 1909. I put it to one side while I did a Win 10 install on a new machine, but that wouldn't install fully either.

Turned out it was my keyboard, swapped it and all was good. So, yes, peripherals.

 

[britechguy]

Turned out it was my keyboard, swapped it and all was good. So, yes, peripherals.

Talk about a situation where, afterward, I'd want to pick up the machine and hurl it into a brick wall or off the top of a building! And the fury is really aimed at Microsoft, as there should be no reason that a bad keyboard or an attached thumb drive or SD card or any external storage should be able to cause a feature update to fail.

Admittedly, what's being described is rare, but whatever it is that is triggering the feature update to fail based on this needs to be tweaked such that it would be ignored, as it is, no pun intended, entirely peripheral to the task at hand.

 

[Sky-Knight]

Drivers... they've been the Achilles' Heel and yet greatest blessing for the Microsoft ecosystem since the registry and invention of the HAL enabled them.

Linux doesn't have drivers, it has kernel modules. That is hardware supporting code that's actually compiled WITH the kernel. So you have version of a module, but it's then further modified by the kernel in question. This approach is more stable, but it's HORRIFIC in terms of proliferation.

Say driver version 1.0 works on three different major kernel revs, and two minor revs per major. That's SIX different kernel modules to track!

NT gave us a HAL, so drivers simply interface with that HAL. And that's how we wind up with drivers that work for Windows 7, 10, and all the server releases and others between.

The problem is... yes you can have a keyboard driver that's no longer compatible... and splat goes some process somewhere that doesn't know it needs to check for it. Because there's simply no way for Microsoft or anyone else to be aware of all the possible options in this relatively open ecosystem.

USB storage is particularly snippy because it's hard to know what storage device is the one you need on a low level. This is double true of idiot mainboards that have terrible BIOS's that change the order devices are enumerated in at random.

This is like Row Hammer... less a bug and more a reality of implementation. Neither will ever be "fixed", unless we can invent a much better methodology.

 

[britechguy]

If you can explain to me how drivers are playing in for something like a keyboard that has been working just fine for years, I'd like to hear it.

What came before your last message does not suggest, in any way, that it's a driver issue for the peripheral. That's why it's so maddening. If I've had the same connected USB keyboard (which most systems do) for years, and it's worked with Windows 10 for years, the last thing I'm going to suspect if it decided to suddenly stop working is that the driver is at fault, and even if the keyboard were to have been snapped in half during a feature update applying that still should not be something that prevents it from doing so.

 

[Sky-Knight]

I haven't seen a keyboard actually cause an upgrade fault, BUT because it's a driver in the device list... yeah in theory it could.

If you can open up device manager and see it in that tree, it can cause a problem. But I find most upgrade issues are storage related, largely due to aforementioned insane mainboards that can't make up their minds which order they're going to present the drive to the OS.

For our fixed storage, the volumes all have GUIDs now, so the OS has something to use to identify it. That's also why BCDEdit is so much more of a chore than XP's old ARC paths for booting. You can't just tell the system... hey... boot to drive 0 partition 1. Because the stupid mainboard will replace Drive 0 with whatever it wants!

That's why on feature updates you get boxes that get to that first reboot and fail... the USB device was in the way somehow, and the easiest way to sort the problem is to remove it.

Keyboards being an issue... I'd love to know specifically WHAT keyboard was involved there, because I suspect it was a compound device and not just a typical USB keyboard running on the stock Human Interface Driver that most use. This is the fault that some printers will cause when they have integrated memory card readers... it's basically a USB stick that's always connected but not in use. Things got SO MUCH BETTER when printers largely got off USB and went wireless.

But again this is the nature of the NT kernel, and all the OS's built on it. We've been doing all this subconsciously for decades.

 

[wavey]

It was an old Dell keyboard, I've been using it for 10/12 years and it was old when I got it. It's in the recycle bin now and has been replaced with another old Dell, similar vintage. I get lots of old mice, keyboards, cables etc from my old regulars.
What was wrong with it? Dunno, I could still type with it but I couldn't install win 10 on a new build, the same spec I've used for a couple of years.
Goodbye old friend.

 

[nlinecomputers]

Talk about a situation where, afterward, I'd want to pick up the machine and hurl it into a brick wall or off the top of a building! And the fury is really aimed at Microsoft, as there should be no reason that a bad keyboard or an attached thumb drive or SD card or any external storage should be able to cause a feature update to fail.

Admittedly, what's being described is rare, but whatever it is that is triggering the feature update to fail based on this needs to be tweaked such that it would be ignored, as it is, no pun intended, entirely peripheral to the task at hand.

And how could that really be done. A faulty keyboard could simply be flipping some random bit in the address bus line. It’s wild but not really something you can harden your software install routine against. This just sounds like bad luck. I’m just amazed that he found it.

 

[nlinecomputers]

It was an old Dell keyboard, I've been using it for 10/12 years and it was old when I got it. It's in the recycle bin now and has been replaced with another old Dell, similar vintage. I get lots of old mice, keyboards, cables etc from my old regulars.
What was wrong with it? Dunno, I could still type with it but I couldn't install win 10 on a new build, the same spec I've used for a couple of years.
Goodbye old friend.

What led you to look at the keyboard?

 

[DRPCNZ]

any USB devices plugged in. Been a few issues with that particularly storage devices, including printer with there card readers. Nothing in the card reader just a bare drive but once the printer was unplugged life was rosey

 

[wavey]

What led you to look at the keyboard?

The windows upgrade and the new install only had a monitor, mouse & keyboard in common, second attempt on the new pc I used a different mouse & k/b, it installed fine. Old k/b & new mouse on the upgrade failed again so changed k/b and it worked.
It might be coincidence of course, but trust had gone.

 

[Sky-Knight]

The windows upgrade and the new install only had a monitor, mouse & keyboard in common, second attempt on the new pc I used a different mouse & k/b, it installed fine. Old k/b & new mouse on the upgrade failed again so changed k/b and it worked.
It might be coincidence of course, but trust had gone.

Oh so you didn't actually narrow it down to the K/B, it was one of those two devices. Still... dang... that's well down the rabbit hole!

 

[YeOldeStonecat]

Just noticing the "keyboard driver" discussion. I can say, at least...two maybe three times in my career...I have seen a computer halt during bootup due to keyboard driver issues. Hasn't been in quite a while. First time was back on an NT 4 Workstation. I just remember that time well, the client, their pretty hot in an Elvira way "lady in charge of the IT over there", etc. Anyways ...yeah. Stopped it cold during a reboot. Probably the one out of the two times in my entire career that choosing "last known good hardware configuration" during a bootup worked.

One would thing that, since you can boot up most computers "headless"...(also implying keyboard/mouseless)....that a keyboard driver wouldn't stop a computer from booting up...but...I've seen it happen a couple of times.

As for why this particular computer isn't completing the upgrade...even when manually running the Win10 creators tool update tool....after a while, just have to throw in the towel and "nuke 'n pave". What other ideas....uninstall antivirus?
Clone to another hard drive? I have seen some upgrades fail...and, was on a spindle hard drive (no sane person should play with W10 on a spindle...but some customers...)...anyways, cloned to an SSD and ran the upgrade again and it flew through. Perhaps some mild drive corruption.

 

[wavey]

Oh so you didn't actually narrow it down to the K/B, it was one of those two devices. Still... dang... that's well down the rabbit hole!

Yep, just a guess, I didn't know what was going on but I had to get the new build out same day. I came back to the upgrade next day.