the computer shop hasn't figured it out yet and now they are closed for the weekend. maybe I should buy another nvme at the other computer shop (no service but open on weekends) and put it in my gaming rig and try to set up Debian on it. Just one slot so I can't refer to my old files yet.
Discussion
What’s going on
My computer has been randomly resetting more and more often for months, always the kernel displaying the same hardware error which was very much incomprehensible (and nothing on google search that matched). It got to the "10 times per day" level and I can't take it anymore.
So I bought a new (motherboard+cpu+memory), plugged them together, and it doesn't POST (no beep, no bios, nothing, just fans spinning).
I bought these new parts at 2 different local shops (no shop had them all in stock). Since I can't tell what is wrong I don't want to just return the motherboard without knowing that is the problem. So I've got one shop testing it on their bench. They said "$79 for one hour on the bench" and I said "ok" meaning in my mind they would figure it out in "one hour"... but I now think they meant $79/hour for however many hours it takes.
Also, I left an old NVME on that motherboard that has nothing installed on it (well it has RISCV stuff that will absolutely not work I was planning to wipe it) and now I'm waking up in the middle of the night thinking they might have been trying to recover my data! I hope not.
Turns out they are open on Sat at 10 AM so in 9 hours (It's 1AM now) I'll go in there and get an update maybe stop them... they have had it 1.5 days already.
In my OP here, I was wondering if in the mean time I could get a jumpstart on installing a new operating system if I buy a new NVMe drive and do the install in a different computer and then move it.
And finally, I have no idea how you guys use phones for nostr. My finger presses 3 of those letters at once, always types the wrong one, and keeps miscorrecting what I type, plus my arm gets sore holding the phone up all the time.
i feel that pain of the hate for touch screens
also none of these silly people can touch type, that's why they don't mind so much, although some do note the epic fail of touchscreen keyboards and autocorrect
standard troubleshooting method is unplug everything then plug back in one by one
if a motherboard shows no signs of life with nothing but CPU installed then it is likely dead
paying somebody to troubleshoot a product within return period is insane
just fucking return it Jesus Fucking Christ !
i mean it could be the PSU of course but PSUs are interchangeable so you can borrow a PSU from another PC and try it on both
if it doesn't work on both PSUs then it's not the PSU
nostr:npub1acg6thl5psv62405rljzkj8spesceyfz2c32udakc2ak0dmvfeyse9p35c
It’s very odd that the shop would not be able to troubleshoot this within an hour. The standard procedure for your error is to unplug everything and reseat it with minimum plugged back in. You’d test for CPU, PSU being plugged in correctly (24 pin especially), RAM (one stick at a time in different slots), and finally the board itself.
Literally should take 30 mins or so to do this if you have a working spare computer.
My bet would be on memory issues - those are usually most common reasons for not POSTing but could be video too. Motherboard and PSU are rarely the issue - especially if new. Memory often comes faulty even brand new.
I had a similar issue not long ago and it turned out to be the 24pin CPU cable not plugged in cleanly. I snapped it in finally and the issue was resolved.
Going for Debian Stable?
Yes for now. With backports if needed and /usr/local compiles if really needed, and even then I may just switch to testing instead. I'm used to Arch which is pretty bleeding edge and it was never much of an issue to be on the newer stuff.
Well, it depends on your preferences. I guess you could live with debian stable with its rock solid base. If you need some updated packages, you could enable the backports repo.
Otherwise you could go for Debian Testing, if that suits you better.
Rough, I had an nvme die very recently too. Didnt even have it long either. Not the first time. Never had this problem with sata ssd's
It wasn't an nvme dying, I typed the details here: nostr:nevent1qqsxrapt685kncrqdwx0ac4l03fzm8gujvye708njrmqtef3q6tgwqszj0lgx
i have an NVMe currently that rapidly overheats, but it otherwise is working, possibly can fix it by ripping off its stickers and putting some good heat goop on it and a heatsink but i just try to avoid doing big transfers
Nothing is wrong with the NVMe drives. People are misunderstanding my OP.
chances are you are using a dogshit chassis with no air flow.
i use Lian Li V3000 Plus. worth every penny.
of course an NVME needs a heatsink but a good NVME like Samsung will have its own copper heatsink already installed when you buy it. beyond that all you need is good case airflow.
a good motherboard will also supply heatsinks for your NVMEs as well.
if you didn't get a heatsink on your NVME or Mobo and it overheats throw them both out and throw your case out as well. throw your entire garbage PC out in fact and buy a Dell.
you are not cut out for this game.
yes
Random panics again?
they got so frequent that I tried swapping hardware, now even new hardware isn't working. at least nothing has exploded.
Damn, that's really annoying. It doesn't seem a fun weekend plan :(
Have your mathematically excluded the software culprit?
The new hardware isn't working. The BIOS is not coming up. Not quite dead as the fans spin and RGB lights up. I haven't installed anything on it yet.
:(
While I wait for the new hardware BIOS upgrade, I took the NVMe out of my gaming rig and installed a new blank one. Today I setup debian on it. Then I installed my workstation NVMe alongside it and also my 2 SATA 6TB HDs with btrfs so right now I have access to all my data again. But on a shit monitor (yes my gaming rig has a small monitor) and at a shit desk.
I decided to make this harder than it needs to be by building new instead of just booting my old workstation NVMe ... every 5-10 years I find it necessary to start over else you get buried in broken cruft and inexplicable behavior.
Every 5-10 years I also prefer to start from scratch, without any written guidance, and I do not use my old dotfiles, but copy and paste the specific parts I need when the need arises.
This helps me to have a fresh system and not to stick to old habits, but to improve them creatively.
Sometimes it's good to start over.
Yep. I've brought over dotfiles as I need them: .ssh, .gnupg, .git, .mozilla (but cleaned it up), .emacs, some .bashrc, and of course .local/share/gossip
Others are easily regenerated by their software: .cargo, .rustup, .cache, .config/* (rarely does any of that config matter), etc.
in US if you get a business class Dell machine you get free next day on-site service. my laptop had problems the next day a tech came to my home and completely took the thing apart and kept replacing everything including the motherboard until it worked. all for free.
Dell doesn't like to use AMD though, but they occasionally succumb to customer pressure and do offer AMD chips like Threadripper which had no Intel equivalent at the time.
I think people under-estimate the value of service. frauds like Gamers Nexus make a huge deal of Dell over-charging people $2 here and there but then downplay problems like other companies delivering computers completely shattered in pieces because they can't engineer a GPU support bracket or proper packaging.
if you browse specifically forums for system admins ( who manage thousands of machines ) 90% of them recommend Dell to each other. but dumb kiddies on Fediverse think Dell is the satan and unironically refer to Gamers Nexus as "Tech Jesus"
Gamers Nexus complained that Dell fan mounting mechanism is "unnecessarily complex" ... what a crock of shit. Those fans are screwed into brackets ( that snap into place ) instead of screwing directly into case and Gamers Nexus said those brackets are unnecessary. But those brackets allow the tech to swap the fan in 5 seconds versus 5 minutes of looking for a tiny screw somewhere inside the case after you drop it on the mobo and short it out.
i hate scum like Gamers Nexus and nostr:npub1kyk7ac33apd7cx0nun3laevf84zfhr8pt8kj4h8v7cpx9t72d4gqkyea0g who make everything about their narrative while ignoring the obvious truth.
the narrative is that DIY is best, small companies are second best and market leader Dell is the worst ... i mean that's why all the system administrators use it - because it's the worst. anybody who believes that is a fucking retard.
of course i understand us cool guys like to build our own stuff.
here are my thoughts of what brands are good for components in a DIY build:
https://dissidentsound.discoursehosting.net/t/hedt-build-phase-1-lessons-learned/728
TLDR: use Asus, Corsair, Noctua, Lian-Li
actually one possibility is that you didn't plug in CPU power connector which is separate from the 24 pin ATX power connector.
make sure every molex power connector on motherboard has something plugged into it.
but i think usually a motherboard will beep when that is the case.
try swapping PSUs. use one RAM stick at a time. try different sticks in different slots.
do not connect anything not absolutely essential to get to POST.
if it is still dead sent it back for refund and get an Asus.
that said i'm sure the shop will have tried all of this.