Windows 10 - Venting

[Creeper]

This is a long story but after a frustrating day dealing with Microsoft Window there may be some useful information in here for some of you.

Two days ago I noticed one of my PCs which I use for a server on my home network wasn't working. The fans were spinning but there was nothing on the monitor. It is a Windows 10 PC I have had for years. My first reaction to problems like this which I have seen before, is to power down the PC, turn off the power supply for about 40 seconds and reboot. That usually solves most issues when Windows gets confused and stops working. This time, Windows 10 started loading but then a message popped up on the screen that a bootable device could not be loaded and the PC would automatically restart (because even Window 10 knows that rebooting solves most of it's horrific code bugs.)

When the system rebooted it went into repair mode - the dreaded black hole of spinning dots with absolutely no information provided to the user. This usually takes anywhere from 5 minutes to an hour depending on the speed of the processor. Of course, this process did nothing and told me nothing. It would have been nice to tell me what device was having a problem because then I would be able to troubleshoot that one devices - or even remove it. I knew it could not be the primary HD where Windows was installed because it started to load Windows from that drive before failing, plus it is an SSD with no moving parts so it is unlikely that was the problem. After running diagnostics (provided by Microsoft) it suggested reloading Windows 10. After hours of frustration I did just that. Of course this means losing all programs and settings, but since this is a server there was not much on it. The process supposedly saves files but I found out later that isn't exactly true. If saved some files on the primary drive up to 2022 but anything since them was lost.

Reloading Windows 10 did not work either. Windows installed but then on reboot I got the same error message that it can't load a bootable device. Out of frustration and a last resort I decided to clear the CMOS and try again. That worked. The system rebooted and I was immediately able to find a problem. Clearing CMOS turned off RAID mode for two 6 TB HDDs I use for file storage, backups etc. But only 1 of the 2 drives was showing in file explorer. After checking physical connections I looked at the status of the 2 drives and the "dead" one appeared unallocated. Since it was RAID 1 (mirrored) I thought I could reallocate the drive, format it and load RAID again. the good drive would copy itself to the mirrored drive automatically and everything would be back to normal. So I reallocated and formatted the drive, then rebooted and reset the BIOs to turn RAID back on. When the system rebooted I got the same error message as before that a bootable device could not be loaded. This is when I realized the error message is a generic message that means nothing. A RAID array for storage is not a bootable device and I have had RAID drives fail many times without rendering the system unbootable, although this was a first in Windows 10. Question: Why not include information about the device that is failing to load? The system must know what it is trying to do when it runs into a problem. Even the name of the driver it can't install or that is not working would have tipped me off to the problem immediately. Microsoft is awful this way.

After trying for hours turning RAID on and off and not being able to find the problem I kept digging only to find a weird solution. Boot the system with RAID mode off, set the PC to boot in safe mode then restart turning RAID mode on before it reboots in safe mode. Why? Who knows but apparently this tells windows to reload the RAID drivers! Again, why? Welcome to the underworld of Windows and Microsoft. This worked but I had to reallocate both drives and reformat them losing all files on the drive. Luckily, at some point I thought it might be a good idea to backup my mirrored drives as one additional step of redundancy. I was able to recover most of what I lost. This is kind of stupid because the entire purpose of RAID 1 is redundancy. If one drive fails, the other operates the same without mirroring and no data is lost. Then replacing the bad drive is a simple process and mirroring works automatically again.

I have no idea what caused this problem to start out of nowhere. The system logs showed nothing. I have never found the Windows system logs to be helpful on system crashes. Most o the time all you get is a message that the system shut down unexpectedly and rebooted. Gee, thanks Sherlock! I never would have figured that out from the blue screen I just witnessed!

In any case, if you use RAID in Windows 10, remember this process if you see a problem like this. Boot into safe mode with RAID off. Then reboot, turn on RAID in the BIOS and then reboot without safe mode and the problem should be resolved.

And don't forget to back up your RAID arrays, even RAID 1 because Microsoft has no idea how that is supposed to work!

 

[Runwildboys]

After reading the first paragraph and a half, I thought I was gonna finally learn how to fix my old laptop, but when you mentioned CMOS, I realized this is way over my head.

I replaced that laptop with a Chromebook, but kept the old one for ripping CDs and syncing them into my phone. That doesn't usually cause too much aggravation, and when it does, a restart or three usually "fixes" it.

 

[CyberB0b]

This isn’t a windows problem, it’s a RAID problem.

 

[Tabascocat]

I don’t even know what RAID is……

 

[Creeper]

This isn’t a windows problem, it’s a RAID problem.

Well, yes, it was a RAID problem somehow caused by Windows. My real issue is really the lack of useful information Windows provides when it craps out. If it can tell me a bootable device cannot be loaded, why not tell me which device that would be? And if it is not imperative to the operation of the system, then why not disable that device and start Windows without it so I can diagnose the problem with the system running - which is much easier. The system was running, then somewhere along the line it black screened. From that point, it would not reboot. Earlier versions of Windows were better at this stuff. I had RAID drives fail multiple times under Windows 7 and XP and the system kept running as designed. And the funny thing is, the drives in the RAID array all checked out on diagnostics and are currently running without issues in the restored system. It does not appear to be a hardware problem.

All I am asking is better diagnostics from the OS when stuff goes wrong. Getting no information, or false information from the OS is just frustrating. I might was well go back to Ubuntu.

 

[Reverend Conehead]

I don’t even know what RAID is……

There are a few different acronyms for RAID, but the one I learned is "redundant array of inexpensive disks." Some years back, they changed "inexpensive" to something else; I forget which. In short, it's using multiple hard disks as if they were one big drive, but with redundancy, so that if one drive fails, the system will still run (though maybe not as well), and so you can simply replace the bad disk into the RAID system, and then the other disks bring it up to speed. The idea is to protect data without having to run backups all the time, though sometimes people will run backup systems anyway so that you have even more redundancy. It sounds like something has somehow gone very wrong with his RAID setup, though I've never set up RAID or used it before, so I'm not certain what.

 

[Runwildboys]

There are a few different acronyms for RAID, but the one I learned is "redundant array of inexpensive disks." Some years back, they changed "inexpensive" to something else; I forget which. In short, it's using multiple hard disks as if they were one big drive, but with redundancy, so that if one drive fails, the system will still run (though maybe not as well), and so you can simply replace the bad disk into the RAID system, and then the other disks bring it up to speed. The idea is to protect data without having to run backups all the time, though sometimes people will run backup systems anyway so that you have even more redundancy. It sounds like something has somehow gone very wrong with his RAID setup, though I've never set up RAID or used it before, so I'm not certain what.

"idiotic"
"infected"
"I don't care enough to come up with more 'I' words"

 

[The Fonz]

Why windows for a server? better use Linux or BSD.

 

[Reality]

Why windows for a server? better use Linux or BSD.

For servers I prefer linux as well, but in this case, the bigger issue is the RAID itself.

Unless you are running RAID1 or RAID10 (mirroring), I would not rely on any "onboard" based RAIDs.

Meaning, I prefer hardware RAID controller card supported RAID arrays because those operate outside of the OS.

It's even better if the RAID controller configurations are stored in the RAID drives because that means even if your RAID controller card dies, you can replace it and your RAID arrays should be intact, assuming there was no physical data corruption when the previous one failed of course.

You would think after all these years that Windows would have improved considerably, but it's still just as clunky and unreliable as it ever was. Even worse, it still requires you to Google just about every error, log entry, etc. to decode what the error is.

The main foothold that has kept the bulk of people I know running Windows are games and outdated-but-still-required applications.

With gaming though, Steam and the Linux gaming community have been making huge strides toward supporting more and more Windows games, and the more that happens the less people are going to be willing to put up with Microsoft's clunky products and privacy-invading telemetry and tracking tactics.

A lot of those old apps are being replaced by web-based apps and services though, so the burden is no longer on the linux community to replace old obscure-but-required Windows apps.

Microsoft should have skipped Windows 11 and focused on building a new Windows OS from scratch without all of the overhead.

 

[cald0d30s0]

Just use Debian or any RHEL OS for servers. Much more stability and documentation.

 

[BrAinPaiNt]

Not sure if it would help or not but have you considered switching from a normal windows version to a Microsoft Windows Server software?

Frankly Linux is the best for servers but if you are not used to linux and want to stick with Microsoft windows...you might be better off using one of the Windows Server software.

For a standard home server the basic windows can work but if you are adding raid systems or more you might want to move up the software that is better designed for servers.

 

[Creeper]

Why windows for a server? better use Linux or BSD.

I was running Linux (Ubuntu) for a while. I converted it to Windows 7 when I was crypto mining. I wrote some software to monitor and control 5 mining machines that were running in my basement using Visual Studio from the server. I don't mine anymore so I suppose I could revert back. Now the server is just for file sharing and backups.

In any case, the system could have been any Windows 10 system. My point was about how ridiculously dumb Windows has become and how their diagnostics do nothing, tell the user nothing and just make it harder to diagnose problems when they occur.

 

[Creeper]

Not sure if it would help or not but have you considered switching from a normal windows version to a Microsoft Windows Server software?

Frankly Linux is the best for servers but if you are not used to linux and want to stick with Microsoft windows...you might be better off using one of the Windows Server software.

For a standard home server the basic windows can work but if you are adding raid systems or more you might want to move up the software that is better designed for servers.

Windows Server is probably overkill for my needs. I had a Windows 7 license available at the time so I used it. I have an aversion to buying anymore software from Microsoft. I still have s few systems on Windows 7. I upgraded to Windows 10 on a few systems when it was free. I like Windows 10 overall with just a few complaints. But it seems to me like Microsoft is trying to make their software so "user friendly" they are removing any references to the technology. When problems occur, the messages you get are usually very high level and obscure. "A bootable device cannot be load" if the depth of their messaging. Windows 10 works great under ordinary circumstances most of the time.

I liked Linux too. It has its own quirks which are slowly getting worked out of the system. I haven't used it in a few years so I have no idea what improvements have been made, but at the time I thought the system itself was pretty solid. The apps written for Linux were a bit more clunky. Of course the choice of apps at the time was nowhere near the variety available for Windows.

At this point I doubt I will do anything. I got the system working again and I am looking at software that does a better job of monitoring and logging disk status. I was running a monitoring and diagnostics tool from Western Digital but it did not report any problems at any point until the system failed. And after I restored the system, the diagnostics logs were removed. This is another problem with Windows.