Dead SSDs

dougalg · 2019-12-14 06:14:03

Hi everyone.

I got Arch up and running on a ~5 year old computer about 2 years ago now. I didn't use it that often, and I had a lot of trouble getting games to work. The system would reboot suddenly when I tried playing games but not consistently, and I figured it was just driver issues, but I never had time to look into it clearly. Then, about a month ago it stopped recognizing the main boot drive (SSD) after a crash. I figured maybe it's time had come, bought a new one, re-installed arch, and got things running, but then today it crashed again, and the new drive is now not being recognized by BIOS at all.

I'm wondering if this is maybe a bigger issue that just graphics driver problems? Maybe I have a more major hardware problem on my hands? I'm not very experienced in dealing with these sorts of things, so I'm looking for any suggestions on next steps. In the past I wondered if maybe I had a bad power supply, or possibly something else. I'm almost wondering if at this point it makes sense to replace the entire machine more than trying to debug it?

EDIT: I booted into an arch USB, and I see a bunch of errors like this

acpi bios error could not resolve _sb.pci0.sat0

Specs are:

Motherboard: Gigabyte GA-Z77X-UD5H
Graphics: Gigabyte GTX 670 Windforce
Ram: Corsair 2x 8GB
HDD: 2 x WD 500GB
SSD: Crucial MX500
Power Supply: NZXT 750W
Processor: Intel i5

Thank you for any help

Last edited by dougalg (2019-12-14 06:30:25)

seth · 2019-12-14 08:18:50

All boards have ACPI implementation errors, usually not critical.

Is the SSD available when you boot from the installation ISO?
What about the old SSD?

Have you checked whether there are BIOS/UEFI updates for your board?

dougalg · 2019-12-14 09:54:33

Thanks, Seth.

The BIOS is up to date with the latest stable version, but there are several beta versions available. But it doesn't sound like they provide any bug fixes.

The SSD is not available when I boot from the ISO (at least not according to fdisk), same for the old SSD.

graysky · 2019-12-14 10:18:15

Boot to the ISO and post the entire dmesg output (you can save it to a USB stick),

dmesg > log.to.post

seth · 2019-12-14 12:49:32

Did you switch the cables (power/sata) when you replaced the SSD?

dougalg · 2019-12-15 01:54:30

Thank you both for the replies!

Graysky, here is the dmesg log: https://drive.google.com/file/d/1sv8ZrN … sp=sharing

Seth, I switched both the power and sata cables when I replaced the SSD.

seth · 2019-12-15 09:10:57

[    4.435432] ata1: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
…
[    9.492628] ata1.00: qc timeout (cmd 0xec)
[    9.492637] ata1.00: failed to IDENTIFY (I/O error, err_mask=0x4)
[    9.805775] ata1: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
[   19.945950] ata1.00: qc timeout (cmd 0xec)
[   19.945958] ata1.00: failed to IDENTIFY (I/O error, err_mask=0x4)
[   19.945961] ata1: limiting SATA link speed to 3.0 Gbps
[   20.259124] ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 320)
[   51.092613] ata1.00: qc timeout (cmd 0xec)
[   51.092621] ata1.00: failed to IDENTIFY (I/O error, err_mask=0x4)
[   51.405787] ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 320)

The drive is connected but doesn't respond. Does it show up (in BIOS when rebooting after the change) when you set the SATA mode to IDE? (What's the current setting, AHCI or RAID? Did you play with that setting in the past?)
Also simply try a different SATA port.

It's weird that two (new? this isn't an ebay drive?) SSDs fail after short usage (which is why my money would have been on a loose cable) and after some crash™
Do you overclock the system?

Also try to strip it down.
Remove the WD disks and the DVD drive, the nvidia GPU (there's an intel IGP), either one of the RAM modules and see whether the disk appears on such bare system (at least in BIOS, you might have trouble booting it)

dougalg · 2019-12-22 08:49:53

Sorry for the late reply, didn't have a chance to play with the machine until today.

SATA Mode:
- Currently in AHCI mode
- Switching to IDE has no effect
- Switching to RAID has no effect

Differerent SATA Port:
- No effect

Overclocking:
- None

The other drive that failed was about 5 or 6 years old, this latest one was new. But they both failed close together in time...

Will try the teardown and post results in a separate message.

Last edited by dougalg (2019-12-22 08:50:47)

dougalg · 2019-12-22 09:09:50

So, when I strip it down, and plug the monitor into the onboard graphics, then I get the beep code for "graphics card error" if I try to use HDMI, but the other ones seem to work OK. Still can't locate either SSD, though.

I'd also love to figure out the root cause of these problems... If I have to buy a new drive, but can avoid it failing that would be OK by me.

seth · 2019-12-22 12:00:55

Try them in (preferably) a different system or maybe w/ a SATA-USB adapter.
Two disks not registering at all is very conicdental, and the universe is rarely so lazy. Notably since there seems to be no issue w/ the HDD and it would oc. really suck if there'd be a problem w/ the board that just snipers SSDs that you keep attaching

dougalg · 2019-12-23 01:12:23

Yes, I think it is either the board, or the PSU. I'm going to try getting a new PSU and installing on one of the HDDs, see if it goes back to crashing or not. If that fails, I'll see about the board too. Could there be a link between sudden system crashes and the drive getting fried somehow? Basically when it crashed, the computer would just suddenly reboot.

seth · 2019-12-23 08:48:09

A broken PSU can cause sudden hard reboots (undersupply) as well as completely fry devices (overvoltage), but there's no way of saying that this is the cause for sure. Otoh I'm not sure whether the dying SSD has simply caused a shortage, inducing the reboot.
(I know this sucks, but broken things do broken stuff and you'd need to hang the devices into a lab to record their behavior)

Computer, Arch. · 2019-12-25 01:35:45

Yeah, post up when you install that new PSU. That is probably the best place to start. I can see a now ~7? year old power supply not being able to give enough juice, depending on how hard it was ridden.

Since you are in the computer already, check the motherboard for leaking, swollen caps.

Memtest86 could also be a step in the near future.

Last edited by Computer, Arch. (2019-12-25 01:40:19)

dougalg · 2020-01-01 06:08:25

So, it's been installed for about a week, and so far, it has not crashed (although I haven't used it too too much, but I have been playing games on it). Memtest86 passed without issues. I will take a look at the motherboard as well. Thanks for the advice!

Arch Linux

#1 2019-12-14 06:14:03

Dead SSDs

#2 2019-12-14 08:18:50

Re: Dead SSDs

#3 2019-12-14 09:54:33

Re: Dead SSDs

#4 2019-12-14 10:18:15

Re: Dead SSDs

#5 2019-12-14 12:49:32

Re: Dead SSDs

#6 2019-12-15 01:54:30

Re: Dead SSDs

#7 2019-12-15 09:10:57

Re: Dead SSDs

#8 2019-12-22 08:49:53

Re: Dead SSDs

#9 2019-12-22 09:09:50

Re: Dead SSDs

#10 2019-12-22 12:00:55

Re: Dead SSDs

#11 2019-12-23 01:12:23

Re: Dead SSDs

#12 2019-12-23 08:48:09

Re: Dead SSDs

#13 2019-12-25 01:35:45

Re: Dead SSDs

#14 2020-01-01 06:08:25

Re: Dead SSDs

Board footer