My current tower started out on Windows, and for some reason after a year or so it started crashing out randomly. Load didn’t matter, it would pass benchmark tests and then crash randomly 5mins after boot. However there was not a single useful error I could find. Installed Fedora, and looked at journalctl after a crash. Immediately I see “GPU has fallen on the bus”. Apparently it is relatively common, but I also found a thread that said it actually can be caused by loose connection. Did a complete reinstall on my GPU, haven’t had the problem again (~6mo now, had both 535 and 545 drivers). Sometimes it really might be a descriptive error message 😆
Fresh from the box OEM hardware, either Dell or HP… I forget. A laptop system.
We couldn’t get the damn thing to do windows updates, which was part of our initial prep for the system. It kept crashing, no useful info from logs.
I booted off of an Ubuntu live image I had on a USB. Turns out, one of the CPU cores on die was faulty. If I reduced the CPU cores visible to the system to one, it worked fine. All enabled? Crash.
A call to support and a quick service visit sorted out the system, but Microsoft’s error reporting was useless at diagnosing the issue.
I’m just trying to figure out why, if you’d programmed it to detect /boot wasn’t mounted properly, instead of having it error out, you’d have it do the install anyway and then snarkily print “hey, jsyk, that thing I just did probably led to system corruption. Hope you’re reading this log and/or feel like spending five hours debugging a nonbootable system! Have fuuuun!”
My current tower started out on Windows, and for some reason after a year or so it started crashing out randomly. Load didn’t matter, it would pass benchmark tests and then crash randomly 5mins after boot. However there was not a single useful error I could find. Installed Fedora, and looked at journalctl after a crash. Immediately I see “GPU has fallen on the bus”. Apparently it is relatively common, but I also found a thread that said it actually can be caused by loose connection. Did a complete reinstall on my GPU, haven’t had the problem again (~6mo now, had both 535 and 545 drivers). Sometimes it really might be a descriptive error message 😆
I had a problem like this with a CPU.
Fresh from the box OEM hardware, either Dell or HP… I forget. A laptop system.
We couldn’t get the damn thing to do windows updates, which was part of our initial prep for the system. It kept crashing, no useful info from logs.
I booted off of an Ubuntu live image I had on a USB. Turns out, one of the CPU cores on die was faulty. If I reduced the CPU cores visible to the system to one, it worked fine. All enabled? Crash.
A call to support and a quick service visit sorted out the system, but Microsoft’s error reporting was useless at diagnosing the issue.
https://pbs.twimg.com/media/DmWLjU0V4AA2IbJ.jpg
It’s soo dripping with snark and I love it.
Whoever wrote that error message doesn’t know how to spell separate.
😳 your username…
More people ought to enable a spell checker in their code editor…
I’m just trying to figure out why, if you’d programmed it to detect /boot wasn’t mounted properly, instead of having it error out, you’d have it do the install anyway and then snarkily print “hey, jsyk, that thing I just did probably led to system corruption. Hope you’re reading this log and/or feel like spending five hours debugging a nonbootable system! Have fuuuun!”
Like why would you do that
The dev spent 1,000,000 hours making the entire codebase their second language and knows it like the back of their hand, why can’t you?