I mentioned on a review for an SSD I’d had some problems that delayed getting the review done. Thought I’d share two separate large issues and what was determined to be the problem.
Lot of text follows, no pictures. Let the eyestrain commence.
The little RAID that wouldn’t
I got an LSI Logic controller card for use with testing SSDs and also because the 3mbit SATA 2 on board my T7500 was making me sad. The card wouldn’t post, I tried it in another machine, also a T7500 (I have similar setups at home and work) same deal, tried one last time in an anciend dead machine, didn’t see anything, figured the card was bad and asked for an RMA.
Next card shows, same issue. This time I dig a little deeper and see that the card, while not showing at POST time, is showing up in the device manager. I go there, check the thing, it’s unable to start, code 10 – description involves ambiguous ID and mentions USB.
I decide to take out my Anker 4 port USB card (common in all systems I tried it on)… bam, thing posts and boots no problem. Odd. I plug the card back in, posts and boots… ok, that’s odd.. I figure maybe something’s wrong, plug in everything USB again, and go on.
Next boot the RAID card doesn’t post… I power things off, unplug the things plugged into the USB card, power it back on, it posts but fails in Windows. I get a new USB 3.0 card.
Anker card + 8tb SCSI USB drive = RAID card won’t init. Replaced with exactly the same card, same issue. Replaced with a Inateck Superspeed 7 port USB 3, works fine. Anker card without the 8TB USB drive, works fine. Wut?
All I can figure out here is the Seagate Backup Hub 8tb (which shows as a SCSI device) + Anker USB card + LSI Logic = something breaks. I did occasionally get an error of command not supported showing up in device manager. Occasionally had the 8tb drive die.
Yes, I failed at removing all things and building back up.
The little Windows that wouldn’t
After the above, I copied off my OS to an SSD that I may be reviewing later on, did a check and everything was good, and had been on the new SSD and the LSI logic card for a few days and everything seemed fine.
Last night I needed to print something – I noticed that it wouldn’t. I quickly googled the symptoms and discovered that Microsoft had pushed an update that killed printing for many – I was one of the lucky ones.
I went into my update history, sure enough, I was the victim of KB4522016… but luckily there were updates available including one that said HP something and I was all like – yeah, fix in the works, let’s install.
It wouldn’t install… said there was an error. Tried a few more times, said there was an error… figured it was time to turn it off and on again.. I rebooted, system came up that I was missing \windows\system32\drivers\ACPI.SYS
Nothing I did would get me past that fail screen. I did a lot including letting the OS cycle somewhere around 8 times to the crash screen.
I booted off a USB, verified the file was not there, realized my USB stick had absolutely no useful tools and wouldn’t let me do anything fun. Plugged in my old SSD (from about 4 days ago,) and compared the system32\drivers directory. There were 216 files fewer in my new SSD than the old.
I poked around a bit and checked the disk (no errors,) checked other directories (no differences,) and determined that nope, 216 files had been wiped. I could find them in the WinSxS directories, and they were on the backup.
I copied about 250 files from the old drive to the new but still no booting. Always missing something until finally I got an error that involved a version mismatch (probably copied an old driver at that point).
I rearranged things and booted off of my 4 day old drive and decided to make a system recovery USB… and was told no. Wasn’t going to happen. Not a useful error message. That part explained in a bit. There was so little different between my last night Windows and my last week Windows drive I didn’t care and just re-imaged the drive.
I decided if I was running out the clock on Windows 10, I wanted it to run out on a copy of my drive and not on the original. I wanted off of the old SSD as fast as possible so fixing it wasn’t a priority, getting a copy made while internet was turned off so no new updates was.
Once I was on the new image on the LSI and newer SSD I did some testing to determine that the SSD wasn’t failing. Didn’t appear to be at least. I wanted to know what was up with not being able to create a boot USB first off and so I set about finding out.
The error was useless – that it couldn’t be done. I looked around and got a hint that this might be due to a corrupted install.
From an admin command, SFC /Scannow said my install was corrupted and it couldn’t be repaired.
Followed that up with an Dism /Online /Cleanup-Image /RestoreHealth and rebuilt the distro image, but got the same error in SFC that there were errors that couldn’t be repaired.
This was concerning because I expected the same issue to happen again when a system update hit. Corrupted install + windows update was what I blamed for the death.
One of the things I read was that when SFC can’t fix an issue it’s time to elevate past local users. I used PSTOOLS (free from Microsoft, no registration, no tracking, just a tool) and elevated from a local admin to a system account with the following: PSEXEC -h -s SFC /SCANNOW
And it fixed it. I went through all the Windows updates, rebooted several times, all good. Created a boot USB from that, no issues any more.