Microsoft veteran says some 'broken by update' PCs were already doomed

It's not me, it's you. Five words that signify the end of a relationship with a toxic partner, or an ill-timed riposte to users tired of broken Microsoft updates. According to veteran Microsoft engineer Raymond Chen, updates from the company aren't always to blame for borked customer devices. Sometimes those devices were …

[JohnSheeran]

Great story but not a new problem. If a corporation is letting their fleet auto update then I guess they are getting what they asked for. Standard practice for the fleet where I work is to do a reboot prior to patching to avoid just this problem. Should it be necessary? Probably not. Is it a fact of life with Windows? Absolutely.

[M.V. Lipvig]

How does that happen? At the very large company I work at, our notification that an update is coming is a message saying the PC requires reboot by X date, at which time it will boot itself. Sometimes I come in from my weekend to find it rebooted while I was out.

[Not Yb]

You have to have an IT department that wants to control updates themselves. These days, apparently some just decide to 'trust Microsoft to get it right' despite the risk, rather than spend the time and salary on testing each new update in an IT test environment before pushing it to all users.

[JohnnyS777]

True that.

Any IT department that is doing its job is going to control the updates. If they don't, they are either incompetent (possible) or terribly under-resourced (inevitable).

Failing to manage updates in the current era of "The customer is the beta tester, for beta or worse" in any enterprise environment is irresponsible.

[Taliesinawen]

IT department don't test updates :o

Not Yb: “You have to have an IT department that wants to control updates themselves. These days, apparently some just decide to 'trust Microsoft to get it right' despite the risk, rather than spend the time and salary on testing each new update in an IT test environment before pushing it to all users.”

How does one totally disable auto updates?

Who could afford a duplicate IT environment for testing?

How would such testing be carried out in a mixed hybrid cloud environment?

[ecofeco]

Re: IT department don't test updates :o

This. ^^

[Not Yb]

Re: IT department don't test updates :o

It's not particularly common outside large banks and big corporations, but it's still at least a theoretical possibility to control updates at a corporate level and possibly test them before deployment.

Microsoft provides tools to control updating at a corporate level, though they're not commonly used by home users. Microsoft Intune is one option, which (of course) requires a license from MS to use. There's also Microsoft Autopatch. I wouldn't be surprised to find other options out there.

[Anonymous Coward]

Re: IT department don't test updates :o

Place I work has controlled updates (anon obviously).

They also have a limited range of devices that employees use (& it's a PITA to try and retain an old device* as they like to have defined small number of hardware options so updates can be tested)

* I have an old workhorse dev machine tweaked to the nth degree, it runs (hassle free) anything I need for my day to day work (which encompasses a lot!) - rules for new machines are they come "vanilla" so needs everything setting up (& no admin rights on your user so have to use elevate privs approach which does not behave nicely with all software installs you need as a dev). So new machine gets gradually setup as & when time allows (its back of the work queue as I'm normally dealing with critical high priority stuff & I am not setting up the new machine just tends not to happen**). Old workhorse was setup when devs still allowed full admin rights.

** Every time I think there's a gap in my schedule to do such a thing, someone high-up spots that gap & dumps something on me! So I'm one of the few people with dispensation to use "retired" hardware that gets Windows updates that have not been tested in advance as need old workhorse as some stuff that new kit cannot deal with as still needs setting up.

[Anonymous Coward]

Re: IT department don't test updates :o

"Who could afford a duplicate IT environment for testing?"

Literally everyone has test environment. The competent ones have also a separate environment for production.

Many people fail to understand that.

[TooOldForThisSh*t]

Re: IT department don't test updates :o

The team I worked with for almost 20 years did. Our responsibility was everything computer related in sales offices across the US and Canada. Everything. We had an area next to our cubicles that contained EVERYTHING a sales office had. Separate from the HQ network. Laptops & desktops. Printers & routers & Ethernet hubs. Exactly the same right down to the firmware versions. A file & print server running the exact same OS with updates & patches every month. The "lab" even had its own T-1 line to the Internet. Any new software or hardware was tested & evaluated over and over before deployment to remote offices. Any non-corporate software or hardware in remote offices was strictly forbidden. If something new was required we spent months evaluating, testing and approving it. We often discovered issues in offices caused by unapproved software and/or hardware.

Now it's all gone. Company got bought and IT got outsourced. No idea how anything works now but I'm glad I'm long retired.

[Doctor Syntax]

Re: IT department don't test updates :o

"If something new was required we spent months evaluating, testing and approving it. We often discovered issues in offices caused by unapproved software and/or hardware."

Did you ever wonder if your months testing might possibly have had anything to do with people installing stuff without getting your approval?

[ecofeco]

Not everyone runs their systems the same way.

M$ is not a uniform system, contrary to popular myth. And it's buggy at every level, and full of CONSTANT unwanted surprises, so every place has to adapt a very ad hoc approach.

Now add in your vendors who have their own junk-ware to deal with on both servers and PCs AND network gear.

Good times. Good times.

[dinsdale54]

Not just Windows.

Back in the day my Unix customers were frequently very proud of the uptime on individual systems i.e. "we can't reboot this system yet, we're nearly at 1000 days uptime". I always pointed out that the time to find out whether your system reboots is not when there's been a power outage and half your server farm is rebooting.

Systems that make it easy to apply patches without rebooting (Unix/Linux) make this more serious, there could have been years of patches & updates applied that are working currently but don't work correctly on restart.

[Jon 37]

Also, even Linux has occasional kernel patches that require a reboot. There are ways to apply many of them without rebooting, which is great because it lets you plan and schedule the reboot. But you should still be scheduling a reboot!

[kmorwath]

"there could have been years of patches & updates applied that are working currently"

No, unless you have a version that allows hotpatching, the fact that the patch was installed but the executables using those files aren't restarted (inlcuding the kernel, but not only that), they are still running the old, unpatched code.

In fact, today Ubuntu asks a reboot quite often - or tells you what needs to be restarted but cound't be restarted automatically.

Because otherwise you jyst have a false sense of security - and vulnerabilities may still be active.

[Paul Crawford]

Re: "there could have been years of patches & updates applied that are working currently"

In fact, today Ubuntu asks a reboot quite often

That will be systemd or dbus in 90+% of cases.

[Doctor Syntax]

Re: "there could have been years of patches & updates applied that are working currently"

Almost entirely systemd. Without systemd it's kernel updates every few months and dbus about twice in multiple years that required reboots. OTOH this laptop gets shut down any time it's not needed.

[An_Old_Dog]

Pending File Replacement

Back in the day (W9X? NT4/W2K/XP?) Microsoft OSes had a thing called "PENDS". If your (or someone else's) installer needed to replace an in-use file, it went into the PENDS queue, where it would wait until the next reboot to be copied to the appropriate place in the file system.

Some person or group developed a way to update in-use code in Linux kernel core/RAM without needing a reboot, but it was a proprietary, payware thing.

[Paul Crawford]

Re: Pending File Replacement

Some person or group developed a way to update in-use code in Linux kernel core/RAM without needing a reboot, but it was a proprietary, payware thing.

Unlike Windows, the Linux/UNIX file systems allow an atomic replacement of in-use files so easy to update at any time. The issue of currently-running code is the next problem:

- Most cases you just restart the daemon, or wait for natural stop/start, and future activity is updated.

- Some key system components (systemd, and occasionally libc) you can't replace the in-use versions as the process is permanent, so a reboot is needed.

- DBUS has the deeper issue of not being able to restart it and preserve existing state, so again a reboot is needed.

The only one fixes I know of is the kernel, oddly enough, as some systems like Ubuntu livepatch can fix some bugs without needing a reboot. The sane approach for a desktop/laptop is to shutdown when you finish in the evening, but for servers that is not so simple. For a private individual you can get Ubuntu livepatch for up to 5 machines for free, we pay for it as a business even though we don't have much more in total.

[An_Old_Dog]

Re: Pending File Replacement

As I wrote, "... it was a proprietary, payware thing."

WAS.

See: https://en.wikipedia.org/w/index.php?title=Ksplice&wprov=rarw1

Also note Ksplice/Livepatch cannot, without reboot, patch some (few?) of the kernel routines.

[Paul Crawford]

Re: Pending File Replacement

No, it can't fix everything. Only certain bugs that the related patches don't alter in-memory arrangements of code or data by too much.

[Sorry that handle is already taken.]

If a corporation is letting their fleet auto update then I guess they are getting what they asked for.

It gets worse. My team has a dedicated workstation to do tasks that our issued laptops can't do effectively. Some of these tasks take days to run. Meanwhile, the company pushes software updates and forces a restart within the next eight hours. We've asked them to add an exception but we just get ignored.

[Homo.Sapien.Floridanus]

Schroedinger’s PC

Schroedinger’s PC: It’s not alive or dead, until it’s restarted.

[Blue Screen of Bleurgh]

I would also maintain full overnight backups where possible prior to any MS update, especially given their unpredictable outcomes.

I realise backups might not be acceptable to some departments running apps 24/7, but better safe than sorry.

Moreover some tin stored in datasheds probably haven't been properly maintained in months or even years, with some potentially on the verge of some hardware failure, dust build-up or any other trivial little thing that has been left to become a big issue when a reboot is needed.

Taking preventative action is again not always convenient for the rest of the business, but an IT dept should be checking the health of all its tins, including controlled reboots prior to major updates, but not just with a day or two gap, but perhaps 4 or 5 days just to make sure that if something does go tits up, it can be dealt with.

And given the huge expense of ram and storage these days, you can bet that a lot of companies (especially the bean counter departments) will insist IT depts make do with old tin and old components while the bean counters cut costs. It's only when those old components can no longer hack it is there the usual blame game

[hoola]

Server, yes, particularly easy with VMs, and should already be in placr. End user devices, no. At best the data in common location and a profile is protected but the actual OS, no. It will only be in vary rare exceptions.

More commonly nothing is backed up, it relys on profiles, drive mapping and OneDrive all of which mean if the device will not boot it is just reimaged.

[Frankly]

Indeed - enterprise server maintenance 101, the 'sanity reboot' process step before you change anything.

Why are such straightforward but important principles so often ignored?

[Anonymous Coward]

Ah yes, that old chestnut: "It was broke before I touched it, I was just trying to fix it, honest officer!".

What they used to call "a likely story".

[MudFever]

I thought it sounded more like the Bart Simpson defence: it was like it when I got here, you didn't see me do it, you can't prove anything.

[FIA]

To be fair, as someone who's fixed other peoples PC it's also quite common.

People don't reboot when they've made a config change, this has caught me out loads of times on my home server (not Windows), make a config change, think it's fine, don't fancy the 5 minute reboot... then when I need it up NOW after it's just crashed I discover the issue.

The other issue is people never ever install software... ever.....

"My computer's not working..."

"And?"

"Well... you were the last person to touch it when you fixed it 3 months ago..."

"3 months ago.. You've not installed anything since?"

"Yeah... well... I've deffo not done anything so you need to fix it..."

<drives over>

"What's this icon? That wasn't there last time??"

"Oh... yeah... I installed <whatever>...."

"<sigh>"

[Doctor Syntax]

"People don't reboot when they've made a config change, this has caught me out loads of times on my home server (not Windows)"

Normally, at least on Linux, it's just a matter of restarting the service.

[ComicalEngineer]

Does anyone else find it strange that Windows versions work much more reliably when the updates stop?

[LBJsPNS]

Not at all.

[A Non e-mouse]

TBF, it's not just Windows that has this problem. How many network engineers have rebooted a switch or router after a firmware update only to find it not responsive because someone didn't save the last config update?

[Groo The Wanderer - A Canuck]

Good old fashioned "blame the victim" slander. :(

[Bran Muffin]

Well, I don't know

As I have mentioned before, none of my 9 computers have ever borked after an update. Windows 11, 25H2, all of them, Enterprise version. No borks after updates and, in fact, no borks after non-Microsoft software updates (except once during an update of an antivirus package). All I ever hear is how "lucky" I am, but that isn't much of an explanation. I do try to keep the computers running properly, and if there is any sign of a problem I address it as quickly as possible, yet even that is an extremely rare occurrence, especially since I moved all the computers to SSD storage.

I'm just stating facts here. If I had problems with the Microsoft updates I would say so.

[IGotOut]

Re: Well, I don't know

Try having a fleet of several thousand. Then come back.

[kmorwath]

Re: Well, I don't know

It might depend a lot on how your fleet is composed. Quite standard Dell/HP/etc. servers and laptops rarey have problems unless you let users install whatever they like. Cheaper PCs with dodgy drivers without updates are riskier. Special machines with less common hardware may be riskier too.

But for the matter I've seen botched updates in *mix appliances too - TrueNAS for example does use the community version for alpha testing as well, it's not uncommon that an update introduces nasty bugs.

[Doctor Syntax]

Re: Well, I don't know

You really shouldn't have said that. You're on borrowed time now.

[Anonymous Coward]

Re: Well, I don't know

Funny, I've experienced more SSD than HDD failures, that one W11 update around New Year that broke sleep, a weird update bug in 2004 that killed audio, W2000 somehow kept constantly failing... 9 boxen on the same feature release isn't exactly a huge sample.

[Locomotion69]

Obligatory xkcd

xkcd

[Rory B Bellows]

To some degree

It can be true. Although in my experience its more often underlying hardware issues, not "Perhaps a new driver" as Chen suggested. A failing hard drive, a bad piece of RAM, or even a finicky power supply that chose this moment to deliver bad voltage. But the user just sees Windows Updates installing right before it all goes south so they blame the update.

It's also true that updates used to be (slightly?) more reliable compared to what is churned out today. In a well controlled enterprise environment it should be easier to conclude who's to blame. But alas, not all enterprise environments are up to snuff, and the majority of home users might not be able to figure out exactly where the fault originated.

[tiggity]

Re: To some degree

The new driver ... that Windows update installs, for dubious reasons*!

* Often some bit of security reason, but often find a "security related" driver update makes things more unstable /worse behaving

[Anonymous Coward]

Idiots never reboot

So where I work, the idiots NEVER reboot, then wonder why Windows is constantly taking a nap. They complain they get weird errors, their machine is slow, they can't get to the internet, the browser doesn't respond, stuff crashes, etc. They get update after update saying "you need to reboot after this update" and they do not. You look at their taskbar, and they've got so much crap running, and so many browser windows open, you can't even see any icons, just a mass of ellipses.

I reboot at the end of the day when I leave, and I've never had a problem.

Now at home, I run Linux and I only reboot on the 1st & 15th when I do cold backups and package updates, but we know Windows memory and resource management is a mess, and it needs to be restarted often.

So yeah, I'd agree with Mr. Chen.

[M.V. Lipvig]

Re: Idiots never reboot

Bull - windows reboots itself after every update, and you aren't given a choice in the matter. The most you can do is delay the update from taking place for a day or so.

[David Haworth 1]

Re: Idiots never reboot

I'm not sure that's entirely true. Yes, Windows reboots during the update, but after the reboot it continues to make changes and it doesn't reboot again.

Disclaimer: I haven't used Windows for a few years.

My Android phone is similar - reboots after downloading the upgrade, then installs it.

[mirachu]

Re: Idiots never reboot

Every update? False. After *many* updates, sure, but not always.

[david 12]

Re: Idiots never reboot

Bull - windows reboots itself after every update, and you aren't given a choice in the matter. The most you can do is delay the update from taking place for a day or so

That's a default setting. I've only ever had a small fleet, 20-30 not thousands, but I haven't had a computer set up that way for ~20 years.

[ecofeco]

Re: Idiots never reboot

LOL wut?

M$ system update reboots CAN be disabled and often are. Or ignored.

13 upvotes? Look at you newbies. LOL!

[tiggity]

Re: Idiots never reboot

In a home environment you have that choice.

In a controlled corporate environment, there will be update management of some form (depends on the company what approach they take), but whatever route they go, there will be some enforced updates & reboots

[nobody who matters]

Re: Idiots never reboot

"....windows reboots itself after every update, and you aren't given a choice in the matter."

errrrr, yes you are given a choice in the matter - you can disable automatic reboot in the settings and require the system to ask first and reboot only when you say so.

The problems created by MS putting out dodgy updates are undoubtedly multiplied by users who don't understand much about how the OS can be configured (which is probably most of them; not because it is difficult to understand, but because they can't be bothered to find out).

[JimmyPage]

Re: Idiots never reboot

Downvoted, because soundbites like that are a magnet for morons.

The true quote is "Idiots never reboot without a plan for failure".

[mirachu]

Re: Idiots never reboot

Speaking of never rebooting, fast boot is a cancer.

[kmorwath]

Re: Idiots never reboot

Windows does not need to be rebooted often - and no OS can't help if an application leaks memory. Still it's pretty stupid to leave a computer on when not needed, and I by far saw more problems with hardware that doesn't get ouf ot sleep mode correctly, than shutting down a PC at the end of the day and rebooting it the next one - today with SSDs even a full boot takes very little, and you have a "clean" system.

And a powered down system is safer than one that it is not. And if the machine doesn't sleep at least, it just burns power for nothing, and many laptop/desktop aren't also designed for 24x7 operations.

[ITS Retired]

Re: Idiots never reboot

The worst thing you can do to electronic devices is turn them on. That is when they most often fail. The power surge, unless designed for, which costs money, can hit design limits in both the power supple and whats being powered. That takes a toll after a while. For electronic equipment it depends on how often it needs to be used. All day, every day? Leave on. If for occasional use, turn it off when not in use. I won't even mention the stress on thermal cycling of components here.

Computer need updates. They cannot be updated when powered down.

[bazza]

Re: Idiots never reboot

Er, I don’t think you know much about electronics or power control in modern computers.

For a start, it’s perfectly normal for a computer or CPU to turn bits of itself off and on just in normal operations without the user ever being aware. And if one is to consider wear and tear on parts, you deal with dopant migration across silicon semiconductor junctions by turning them off as much as possible.

What tends to fail is capacitors and fans. Spinning fans less lengthens their life.

Liquid electrolyte capacitors are a bit odd; they prefer being kept powered up, preferably somewhere close to their operating voltage. Though good modern ones are far better than the old ones and don’t dry out anything like as quickly. It used to be a common design mistake to fit, say, 63V caps then operate at 12V.

I don’t think anyone is using thermionic valves in modern computers.

HDDs are about the only thing I know of that - if old with many operating hours on the clock - can be a bit grumpy about spinning up from cold. I have on occasion revived such drives with a light tap from a hammer to get the bearings unstuck. But if your drive is in that condition it is EOL anyway.

[kmorwath]

Re: Idiots never reboot

Desktop laptops HDDs are designed diffrently from server ones. The latter are designed for 24x7 operations, and to ensure better performance, are always on and park heads only when turned off properly.

Desktop/laptops HDDs are designed to sustain power outages and the like, park heads more often, and are often turned off when not in use. And are usually designed for 8x5 operations.

[Yet Another Hierarchial Anonynmous Coward]

Re: Idiots never reboot

My fix for HDD's that won't spin up after power outage is to hold horizontal and rotate vigourously back and forth a few times to give the disc a bit of inertia. Preferably when unpowered, but occasionally stubborn ones need to be powered and attempting to start under their own steam. That's also the point you know it is fooked and needs to be cloned/replaced without delay.

[kmorwath]

Re: Idiots never reboot

Don't turn off your car, or your oven then. There are devices that are deigned to be always om. and other designed to be turned off and on.

[bazza]

Re: Idiots never reboot

Agreed, though (at least for an internal combustion engine, and especially diesels) they benefit from being used regularly and driven hard every now and then, to get good and hot. Left off all manner of seals dry out, fuids go gooey, etc. A car just left will evolve problems without apparent changes in entropy.

Turns out that car engineers are very good at designing cars for a specific lifecycle / workload, to a price!

[David Hicklin]

Re: Idiots never reboot

> more problems with hardware that doesn't get ouf ot sleep mode correctly, than shutting down a PC at the end of the day and rebooting it the next one

I have always properly shut something down if I am finished with using it, especially after bad experiences with sleep the worst being going from WFH to docked at $WORK never worked reliably.

[Paul Hovnanian]

Re: Idiots never reboot

Yeah. No.

Depends on whether it's a 'server' or a 'desktop'. I've got one Linux system that's been up for 112 days (since a bad windstorm). But my desktop (also Linux) gets booted almost daily (uptime 1 day, 58 min at the moment). Not because it slows or hiccups alot. But just because it's a waste of power to keep it running all the time.

I'd venture a guess that most people shut down and restart Windows desktops pretty often for much the same reason. And if that's the case, most boots occur NOT because of an update. But because you came to work and turned the thing on. That said, I'd suspect that a bad hardware/configuration problem would reveal itself more often than only following an update. And most people would notice the difference. "I turned it on and it shat itself" is different than I turned it on, it came up fine, but then announced it was installing an update. Only then did it go T.U. I think most people can tell the difference.

[Evil Auditor]

There's nothing like the thrill I experience every time after my Windows workstation installed a patch: will it boot again? Will I be able to log in? Will I find my usual desktop? Or will it... fail?

[IGotOut]

Ah yes, the driver fault...

..that'll be the CORRECT driver for the part, that you then rewrote with a downgraded PoS, that you then have to fix by reinstalling the correct driver again. Graphics cards were the most notorious for this

So excuse me while you say "the logs showed you installed a driver, so it broke it" and call it bullshit. After all who installs a driver and doesn't reboot?

[Doctor Syntax]

Providing it will boot off an optical drive or USB it won't be unsalvageable.

[goblinski]

"My colleagues over in enterprise product support often get corporate customers who report that 'Your latest update broke our system'."

From my recent (past couple of years) experience with Microsoft support - I'm not sure Mr Chen is using the correct verb tense here.

Maybe there was a time MS Support was a thing that worked, nowadays it seems to be an endless string of one "engineer" taking over from another, and asking for data that was requested by and provided to the previous engineer hours earlier and is literally sitting a few paragraphs back in the reply thread, bulletpoints and all.

This - on separate cases at separate times with separate "engineers".

[JimmyPage]

WALOB

If a "broken" machine can run fine in Linux then it's Windows - no matter how it was borked.

[billdehaan]

He has a point, but the issue is that the architecture allows it

It's not like this is a Windows specific problem. If you update your Linux kernel without rebooting, then install updated drivers while still running the old kernel, and the drivers configure themselves based on the in-memory kernel version, there can be conflicts when the machine eventually does reboot. And they can be significant conflicts, when the updated drives can't talk to, or find, SSDs, keyboards, monitors, etc.

The difference is that in Linux, you can defer updating the kernel as long as you like, and when you do, most of the update tools will advise you to reboot as soon as possible.

I've worked in Windows shops where many desktops ran for months without a reboot. In many cases, the desktops had longer uptimes than some servers. That was fine when the machine was static, and when the IT department managed updates. But when the users are installing, upgrading, and uninstalling apps, it can be chaos. We had one instance when a machine hung when rebooted for the first time in four months, and the issue was tracked down to one vendor's video update six weeks prior that conflicted with a second vendors graphics library update two weeks after that. Machines that had rebooted after the video driver update and then updated the graphics library after the update had no issue, but there were three or four machines that hadn't, and every one of them failed to reboot.

With that said, the quality of Windows updates over these past few months has been nothing less than disastrous. They've broken disk encryption, user logins, notepad, the start menu, the task bar, paint, and dozens of everything. They've been so incompetent that users are refusing to update because of the chaos. That's going to result in cascade effects until Microsoft gets their act together and starts rebuilding trust.

Given that right after Microsoft announced they were going to stop focusing solely on AI and actually start fixing all the Windows problem, they pushed out a release that they then had to immediately issue another emergency patch to fix it, I don't see Microsoft regaining user trust any time soon.

[david 12]

Re: He has a point, but the issue is that the architecture allows it

The difference is that in Linux, you can defer updating the kernel as long as you like, and when you do, most of the update tools will advise you to reboot as soon as possible.

In Windows, (XP,Vista,7,10,11, home, pro, embedded, enterprise), you can disable updates as long as you like, and you can set Windows to not reboot ever on updates, and if you are doing manual updates (instead of silent updates) the update tool tells you that you need to reboot to complete the update.

I realize that many people who are not Win Admins aren't interested in how their Windows machines are configured, but that's not a fair comparison with people who update their Linux kernel.

[Anonymous Coward]

Re: He has a point, but the issue is that the architecture allows it

" you can disable updates as long as you like"

Only in Pro/Enterprise-versions with corporate licenses. Otherwise pure BS.

[Anonymous Coward]

Re: He has a point, but the issue is that the architecture allows it

"I don't see Microsoft regaining user trust any time soon."

For Microslop it's totally irrelevant: Typically users are not their customers, they are just users who have no choice whatsoever.

Any corporate machine is bought at top level and no-one who knows, is consulted.

Home users basically can't buy a computer without Microslop installed, so hardware makers are the actual customers and they do want to sell MS-products, so windows is sold with every machine. No choice there either.

Users outside of either group are basically nil: No-one buys Windows licence at retail price from a shop.

[Daniel Pfeffer]

Re: He has a point, but the issue is that the architecture allows it

Home users basically can't buy a computer without Microslop installed

Not quite true. Many manufacturers are selling laptops (and possibly desktops, haven't checked) with no O/S, with FreeDOS, or with Linux installed. I agree that the first two options are unlikely to be chosen by the typical home user, but they are available.

[Doctor Syntax]

Re: He has a point, but the issue is that the architecture allows it

OTOH the OP's primary point stands - MS's customers aren't the users.

[anonymous boring coward]

Re: He has a point, but the issue is that the architecture allows it

"That's going to result in cascade effects until Microsoft gets their act together and starts rebuilding trust."

They have had decades to try this.

[tekHedd]

Sure, I suppose. And then there's the rest of us

Systems like my gaming PC, that sat idle for a month because I really didn't have time to game, got no updated anything other than automated Windows updates, and stopped booting Windows 11 after the latest update. Could not rollback or recover. The drives are fine.

It's now running Arch.

[Excused Boots]

Re: Sure, I suppose. And then there's the rest of us

You know I would really love to know the reason why someone had downvoted your post*. You hit a problem which you managed to resolve by.....

Yes OK, you possibly could have fixed the Windows issue given a lot of time and effort and no guarantee, but instead you decided to try another track which appears to have worked for you. So fine.

* is that you Satya?

[kmorwath]

Re: Sure, I suppose. And then there's the rest of us

Becasuse it's a useless post without real facts - the usual linux fan "ahhh I was using windows and I had <problem here>, I solved switching to <distro>".

[anonymous boring coward]

Re: Sure, I suppose. And then there's the rest of us

So it was you. I guess you think everyone just lies then? Projection?

[ecofeco]

He's not wrong, but he's not 100% right either

Having fixed tens of thousands of PCs, I've seen plenty.

While I hate M$ with a burning passion, I've also seen a LOT of neglected and sketchy PCs.

And therein lies the problem: M$ is slop, but far too many people are fiddling their machines into oblivion. And the majority are never doing any updates.

And then there is issue of other companies also borking their own updates.

BUT, M$ is by far the bigger culprit. They created the whole situation.

[Anonymous Coward]

Re: He's not wrong, but he's not 100% right either

Yeah, I try to be somewhat sane with my game box (as much as you can be sane when the OS started its life as Windows 7 and has been "upgraded" first to W10 and then W11 so there's a decade of cruft) and the PC is perfectly stable, as long as the hardware doesn't fail (had to migrate the OS two weeks back when the system drive started to fail). Fast boot is off, box goes to sleep every night, gets rebooted basically when Windows Update thinks it has to (and even then I routinely delay the reboot as much as possible) and/or I update GPU or chipset drivers. Uptimes are generally several weeks.

[ecofeco]

Re: He's not wrong, but he's not 100% right either

Good practice. Cheers.

[Omnipresent]

You're using it wrong.

This is the "pc" (public computer) version of the apple argument. These nerds are such hubris narcissists. "Me, Me, Me, what about ME! I'm special, you're not!"

[Anonymous Coward]

Re: You're using it wrong.

Yep. Apple told us the same thing once - "Major OS updates sometimes reveal existing hardware issues. Your hard drive was bad before the update." Really? Worked flawlessly immediately before the update, afterward ran at <1% of original speed but without any errors (including copying the entire drive to another machine, which took a week), but somehow it's a failed hard drive? More likely it was a bad or mismatched firmware update to the drive, packaged with the OS update, but the "Genius" Bar only knows how to plug in a cable, hit a button, and read the one-word troubleshooting result "FAILED". So of course it's the hardware's fault. (Seriously, the word FAILED was all the result that it gave. No actual troubleshooting took place.)

Then they put in a larger drive than we agreed on, without asking or even telling me when I went to pick up the machine, and tried to charge the price of the larger drive. Absolutely not, I'm not paying $50 over the already-inflated quote you gave me. A manager intervened when I started to make a scene in the store, and honored the original quote. I only discovered the larger size when I got home. (And then the replacement drive failed a month later, but thankfully it was under warranty and they replaced it again for no charge.)

[david 12]

Re: You're using it wrong.

ran at <1% of original speed ,,, but somehow it's a failed hard drive?

Yes, when my drives start to run at <1% of original speed I sometimes move them to another computer to copy off the contents, but I've never had one "come good" that way.

Just using a different driver or a different OS has never fixed "slow drive indicating failure" for me.

I don't include "slow drive over network" as indicating failure. Once I start using open source tools or cross platform connections, "slow" doesn't indicate drive failure.

Tangent: over my career, worst point was around about 200GB. Those drives, several brands, failed much faster than 2GB or 1TB drives.

[Anonymous Coward]

The idea that it was a few weeks before Windows rebooted is hilarious. It never goes that long.

[hoola]

Window can very easily run for months if you don't care about security patches or it is isolated.

Once you add other software into the mix things can become more muddled.

The less interaction the better. I have seen Hyper-v nodes with up-times of a year.

[Paul Crawford]

Window can very easily run for months if you don't care about security patches or it is isolated.

We had an isolated Win7 machine run for years without issue, but then it was before the current slop of updates and was isolated from Internet and users, so not monkeyed with by anyone.

So yes, an unmolested NT kernel is fairly good, but no, a typical Windows machine these days is not.

[ErikOnTech]

The cope is strong with this one.

Windows updates have always carried a risk of breaking things, but the trend has been greatly pronounced over the past year or so.

Whether or not this is related to certain rather extreme changes in coding practices enforced by Microsoft during that period and several months before, is a question that each of us can ask ourselves.

[bill1977]

Microsoft and bloated software

Microsoft has always produced over-engineered bloated software. They have segued into over-engineered, AI-assisted, bloated, spyware that makes 1,000 connections to the internet on every keystroke.

Does their software need to be updated and tested before takeoff. Sure. Both things can be true - the need for updates/testing and a pile of garbage bloatware.

[Doctor Syntax]

Re: Microsoft and bloated software

This is a use of "over-engineered" I've not previously encountered.

[Bebu sa Ware]

"an earlier era of Windows when Microsoft thoroughly tested code before shipping it"

I am pretty sure the History Monks edited out that bit; probably in a vain attempt to keep things credible.

At this point I can foresee much of the 21st century landing on the cutting room floor.

[Pascal Monett]

Blame the customer

Always a good idea.

[SonWon]

Hah hah hah

Hah hah hah, killed my last Windows PC with Bazzite. No problems with Windows update now. I will fire up my Win 11 VM sometime next month and update it. I only need that VM once in a blue moon for two pieces of USB legacy devices that need updating so there is no hurry.

[Anonymous Coward]

I call BS

Given that anyone using Windows will shut down their machine at the end of the day (on account of it not being very good at sleep or suspend) I don't see that one fly.

Nice try, though, but I'm not buying it.

[anonymous boring coward]

Re: I call BS

Yes, Windows Sleep is a disaster.

And that's AFTER having spent much time sorting out all the device settings where random devices can wake the PC up. An embarrassing mess.

[coisa]

Funny to me

Say you're working along, as an individual, no IT department. You trust you machine to help you make money. You start a windows update cause you don't want to be exposed to a security risk or don't want to restart later. You run an errand and come back finding your computer is bricked by making the update. MS says it's not them, it's your machine.

Yeah, right every successful business knows you blame customer. (Oh yeah, MS says the updates are forced, you must make them)....

[anonymous boring coward]

"It wasn't the update that broke their system. It was the fact that the system rebooted."

Sure. How about a quick reboot then, before applying the update? To prove that things work, or don't work, before the update.

Also, can we not have automatic updates when rebooting? Not everyone has a spare 30 minutes sitting around...

[M.V. Lipvig]

They were doomed

DOOOOOOOMED! Because they run M$. My old machines that barely ran Win7 are ticking happily along on Linux, and it wasn't nearly as hard to learn to use as I was worried it would be.

Mint is a lot like Win7 with one important difference - it just works. Win7 was way preferable to 10 or 11, but still required that I go back to an old restore point every few weeks. No idea what was going on, but it would get sluggish and regular reboots did nothing. Only going back to a restore point from 2017 or 2018 would restore performance. That machine finally physically fell apart and was replaced.