I have a NUC8i7HVK with 2x Samsung 970 Pro 1TB NVME and 2x 16GB Corsair memory from the compatability list. When I use the PC at home, everything this fine. When I disconnect everything, take it to work, use it there, then go back home and plug it in, the OS becomes corrupt and Windows gets in a automatic repair loop.
I've taken the NUC to work twice now, and twice it happened, having me reinstall the PC.
The second time, the recovery console could not access the drive that had the OS on it (recognised the drive letter, could not CD into it).
When I installed Windows 10 the 3rd time, it would show the partitions, but have the drive show completely empty (e.g. 9??GB partition, 9??GB free while 3??GB was used).
I've downloaded and runned the Samsung magician but it reports no errors.
I bought the NUC to replace my laptop and have a powerfull machine to use at work and at home, this is a big problem for me, after spending nearly 2k on a NUC. Please advise.
Is your NUC8i7HVK powered by the same AC adapter at work as it is at home?
Also, are you using RAID with your two 970 Pro NVMe SSDs? And is the BIOS up to date?
Is your Corsair memory capable of supporting XMP 2.0 profile and if so, is it being used?
Yes, I also bring the power brick. SSD's are not in raid, all updates are applied with the Driver & Support assistant.
Memory is "Corsair CMSX32GX4M2A2666C18 Vengeance 32GB (2x16GB) DDR4 2666Mhz 204 Pin SODIMM Performance Notebook Memory Kit", which is on the compatability list for the NUC. I have not changed BIOS settings/overclock.
The NUC functions fine, when it stays plugged in, I have no issues, only when I travel back from work that it dies and needs the OS reinstalled...
Windows automatic repair is often triggered by dirty OS shutdown, e.g. board powered down before Win10 completed a clean shutdown. This doesn't seem to be your case, though.
This is a rather strange case of location sensitivity, so a bit tricky to diagnose.
BTW, are you running Windows 10 1809 (RS5)? There is a nasty issue with the BayHub Tech card reader driver on Intel's website for NUC6i7KYK that induces BSOD crashes specifically with latest Win10 1809. If your OS is Win10 1809 via clean install or upgrade, it is best to uninstall this BayHub Tech CR driver and go back to MS inbox card reader driver.
Yeah, it was a clean shutdown both times.
Windows 10 PRO version 1809 build 11763.195 (Action Pack), installed from USB with media creation tool. I'll uninstall the driver, but had no issues with BSOD's (didn't use the cardreader yet).
It's just such a hassle to reinstall every time:
- Winrar 561
- Notepad++ 7.6.1
- Cisco AnyConnect 3.1.14018
- Filezilla 3.39.0 x64
- LastPass 188.8.131.52
- NodeJS 10.14.2 x64
- Microsoft SQL Server 2017 Developer Edition
- Sql Server Management Studio 17.9.1
- Visual Studio Community Edition 15.9.4
- Tortoise Git 184.108.40.206
- In this case, I suggest you that you should prepare a disk backup. You may do this with software like AOMEI backupper: https://www.backup-utility.com/ (it is Freeware). In case of additional corruption, you will be able to restore your system in less that 1 hour.
- Since you have 2 same SSD drives, can you replace the drives in the M.2 socket? I mean next time when you going to reinstall System, use the other drive.
- I believe that this failure will repeat also if you disconnect Power Adapter during the night at your home?
- Please attach a image of you Disk Management
That would be very annoying since I have my GIT clones on the 2nd drive (500.000-ish files), databases, etc, total ~300GB.
I might try it, but it's such a pain in the *ss :-/
Creating a backup now, will try to get everything into my T5 external SSD drive and leave it unplugged tonight...
Can a depleted CR2032 coin battery be the cause of this issue? Just thinking out loud: if the issue is induced by AC adapter being unplugged from the NUC for some time...
Got to work, same deal again, but now the NUC had trouble even getting into the BIOS. While in the BIOS, I noticed it would see the drive in slot 1 but not recognise it. I powered down, removed the SSD, put it back in, and tried again: no dice. Powered down, swapped SSD's, now it sees the drive that was in slot 2, in slot 1, the 'problem' drive still doesn't show, so I guess it's the SSD. I'll contact Amazon/Samsung to get a RMA :-/ Kind of glad it wasn't the NUC (at least, as far as I can tell), will update if there is more news.
That is the reason I asked you to swap the SSDs. Could be that there is some problem that affects this SSD functionality while you are taking it to your work. May be vibrations? Please post what will be the results of SSD replacement.