X380 Yoga boot issues and crashes

[archetypitect]

My X380 has been having increasing stability issues since December, which this week got unbearably frequent. I have a dual-boot Win10/ElementaryOS, and both exhibit what I assume is the same issue:

Windows 10: regular BSOD, perhaps every 1-3 hours this week. Stopcode: CRITICAL PROCESS DIED. Event manager only shows error 41 (improper shutdown) - not sure if there's anywhere else to look for useful logs. (I mostly use ElementaryOS)

ElementaryOS: Locks up, some applications crash and title-bar icons are replaced with placeholder/blanks for maybe 10s before the whole thing goes down. Error screen: https://imgur.com/d0Fw0ds

Also, when powering on or restarting, the laptop often doesn't make it to grub - it goes to the boot device select screen where the NVME drive doesn't appear. Sometimes a single hard reboot brings grub back, sometimes I do it several times without success. I thought it could be the SSD, but after this started happening on my previous drive (an upgraded replacement) I put the original drive back in, and the problems came back after about a month.

This issue has been driving me crazy - any ideas? The computer is still under warranty, so maybe I should just take it in, but I'd prefer to figure it out myself since I need the thing for school. Thanks!

[dr_st]

Sounds like bad RAM or bad NVMe controller; RAM is soldered anyways, so whatever it is - the motherboard would have to be replaced.

Can you try to run Memtest from a bootable USB drive?

[archetypitect]

Thank you - replaced the motherboard, all is working well now.

[archetypitect]

So I just had the same crash happen again, once yesterday in Windows, and just now in Linux. It's still far stabler than before, but if this has happened on two different hard drives and two different motherboards, I really don't understand what could be happening here. Could a recent BIOS update have borked something? Might be time to reach directly to Lenovo support? I can't afford for this to happen during April exam season.

[RealBlackStuff]

Sounds more like the infamous Thunderbolt problem has risen its ugly head again!

[dr_st]

Sounds more like the infamous Thunderbolt problem has risen its ugly head again!

Question: in the list of symptoms described in the article:

USB-C port not working
Intel Thunderbolt controller not visible in the OS/Device Manager
USB-C or Thunderbolt docking stations not visible or having connectivity problems
HDMI output not available
System battery not charging with a USB-C power adapter connected to the USB-C port
Intel Thunderbolt pop-up error message
Intel Thunderbolt safe mode error message
BIOS Thunderbolt communication error or hang during POST

Where did you see anything that resembles the issues the OP is complaining about?

[RealBlackStuff]

... hang during POST

[dr_st]

He is not complaining about hang during POST. He is complaining about crashes, bluescreens and missing NVMe drive in the boot menu.

[RealBlackStuff]

Agreed, Lenovo has only admitted that the TB3 is/was screwed up, but what about all the other errors that they won't admit?
Unless you absolutely must, you don't upgrade a BIOS, just because it happens to be part of a rollercoaster train of upgrades, that get shoved down your throat at every (in)opportune moment.
And the Lenovo BIOSes of late apparently get "tested" by unsuspecting customers, just because their so-called "Quality Control" no longer seems to exist.
It's not just W10 that screws up, Linux does similar on OP's machine.

Which now actually brings me to doubting the quality of the SSD.
I'd suggest OP gets another (different brand) SSD and reinstalls his OSes.

[archetypitect]

Thanks - I had read about the thunderbolt issue, and there is a thunderbolt update in Vantage that always fails to install, but the symptoms just aren't the same. So yeah, wondering if there's a broken update somewhere, perhaps BIOS, that breaks this config even with new hardware. Since the mobo is new, I'll see if they give me a new SSD. But I don't like the idea of ignoring updates just in case they break the system - it's not exactly secure.