BIOS stuck on RAID ON (INTEL RST) and can't switch to AHCI

[Limpuls]

Something weird happened on my Dell Inspiron 5401 laptop recently. I've been running dual boot Windows 10 and Manjaro Linux for almost a year now on that machine with no problems. Reently I took my laptop on vacation without a charger and it was dead for like 4 days. I also took it through airport security multiple times. When I came back home I was no longer able to boot into either Windows or Manjaro. I get GRUB error: device UUID=*****... not found. Skipping fs ck. You are now being dropped into an emergency shell.
After going into BIOS I noticed that my SATA interface was set to RAID ON. I don't remember what it was before but I only have one built in NVMe drive on this laptop so I'm pretty sure it's not raid. Also it looks like you can't install Manjaro in RAID mode so it had to be AHCI before. But when I try to select AHCI save it and restart, it goes back to RAID again on the next boot. 

I also saw that BIOS version was 1.1.0 which is like from beginning of 2020. I was time to time booting into windows to get BIOS updates so there is no way it could be at that old version. I found this article that mentions similar issue that I have just for a different laptop https://www.dell.com/support/kbdoc/en-lt/000189014/cannot-change-sata-nvme-operation-to-ahci-on-the-xps-9305-ultrabook?lwp=rt and it suggests to update your BIOS. So I did and flashed the newest version. Still no luck. 

Does anyone have any suggestions what could have happened to my system? Is it CMOS battery problem? I can flash BIOS updates and save other settings in BIOS successfully so I doubt it's that. My NVMe drive can also be found and is displayed in BIOS so not a drive issue either? Any help is appreciated, thanks.

[Beerzerker]

Replace your CMOS battery, then reconfigure the system in the BIOS because it sounds like that's the deal here.
If it can't hold your settings that's what it's going to do.

[Pandur]

Raid mode is AHCI for any drive that you haven't built a raid array with. They should work the same, especially with only one drive installed. If your m.2 ssd is nvme and not sata then this is completely besides the point to. A nvme drive is connected to the pcie bus and does not use the sata sybsystem in your computer. My laptop has just one nvme drive and I have completely disabled sata in the bios on that. Works 100% fine.

 

Check that uefi compatibility mode is disabled and that secure boot is enabled (or possibly disabled, not all linux distro are pals with secure boot). Make sure you are in pure uefi mode and try both secure boot on and off.

 

"GRUB error: device UUID=*****... not found" basically means that GRUB was unable to locate the partition with that UUID. GPT partitioned drives have a unique identifer for every partition, and that will change on a re-format/partition. This could indicate that something is wrong with the data on your drive.

 

So either your bios have reverted to compatibility mode instead of uefi only mode. Or your nvme got problems.

[Limpuls]

3 hours ago, Pandur said:

Raid mode is AHCI for any drive that you haven't built a raid array with. They should work the same, especially with only one drive installed. If your m.2 ssd is nvme and not sata then this is completely besides the point to. A nvme drive is connected to the pcie bus and does not use the sata sybsystem in your computer. My laptop has just one nvme drive and I have completely disabled sata in the bios on that. Works 100% fine.

 

Check that uefi compatibility mode is disabled and that secure boot is enabled (or possibly disabled, not all linux distro are pals with secure boot). Make sure you are in pure uefi mode and try both secure boot on and off.

 

"GRUB error: device UUID=*****... not found" basically means that GRUB was unable to locate the partition with that UUID. GPT partitioned drives have a unique identifer for every partition, and that will change on a re-format/partition. This could indicate that something is wrong with the data on your drive.

 

So either your bios have reverted to compatibility mode instead of uefi only mode. Or your nvme got problems.

Hi, thanks for the answer. That makes sense regarding SATA interface being irrelevant for pcie drive, I had that in mind but couldn't find such info onine and BIOS going back to RAID seemed like a weird behaviour. I can't even choose "disabled" option in there. Goes back to RAID ON after restart. Now I booted into Manjaro live USB. But I cannot see my NVMe drive in there. It's not listed like it's dead or not connected. Neither

lsblk, df, fdisk -l

shows it. But I don't it can be broken because I can see it listed in BIOS as boot device and so on. So being able to set it AHCI seems like would solve the issue or not? If not, then do you have other ideas why I'm not able to see the drive in live boot with any of the commands?

[Pandur]

I have access to the manufacturer bios menu on my laptop. I disabled sata there. Like you discovered, usually not an option for that in the regular menu.

 

If the drive is visible in bios it should not be entirely dead. Do you have Windows installation media available? Can get access to diskpart in the troubleshooting section there. Just to see if Windows can detect it at all. And check if there are any intact partitions.

[Pandur]

So, if you have a Windows bootable usb try that. I apologise if you already know the following. Just making sure we're not bouncing messages back and forth unnecessary.

 

When you get to region/language selection just pick your preferred settings and go to next screen. You should see a big Install button in the middle of the screen, ignore that. Bottom left corner has a repair my computer link. Hit that

Then select Troubleshoot - Command Prompt

In the command prompt type diskpart

When diskpart has started type list disk

If any disk other than your usb device is shown, that should probably be your nvme. If so, check size to make sure. There should also be a * in the Gpt column if the boot sector is intact

Type sel disk <number matching what you think is your nvme>

Type list part. This should provide a list of partitions, if any. This is how it looks on my computer:

List disk:
Disk ###  Status         Size     Free     Dyn  Gpt
--------  -------------  -------  -------  ---  ---
Disk 0    Online         1863 GB  1024 KB        *

list part:
Partition ###  Type              Size     Offset
-------------  ----------------  -------  -------
Partition 1    System             100 MB  1024 KB
Partition 2    Reserved            16 MB   101 MB
Partition 3    Primary            194 GB   117 MB
Partition 4    Recovery           517 MB   194 GB
Partition 5    Primary           1667 GB   195 GB

If you can't find your device here either, something is seriously wrong. And by now I do suspect with the drive itself.

 

Imho the next step would then be to test the drive in a different computer if possible

[Limpuls]

15 minutes ago, Pandur said:

I have access to the manufacturer bios menu on my laptop. I disabled sata there. Like you discovered, usually not an option for that in the regular menu.

 

If the drive is visible in bios it should not be entirely dead. Do you have Windows installation media available? Can get access to diskpart in the troubleshooting section there. Just to see if Windows can detect it at all. And check if there are any intact partitions.

Yes I do have win10 bootable. I just got to troubleshoot section and managed to run command line with dispart utility. I typed in

list disk

like you said and I can see my NVMe in there. Status online and GPT selected.

list part command gives me back 3 partitions:
Partition 1 is Type System 300MB
Partition 2 is Type Unknown 300GB. This is my Linux partition I'm pretty sure.
Partition 3 is Type Primary 175GB.

So the drive is alive and all and I can be spotted in diskpart utility. But is not listed as a block device in Linux. What does that mean and how do I proceed from here? 

 

[Pandur]

2 minutes ago, Limpuls said:

Yes I do have win10 bootable. I just got to troubleshoot section and managed to run command line with dispart utility. I typed in

list disk

like you said and I can see my NVMe in there. Status online and GPT selected.

list part command gives me back 3 partitions:
Partition 1 is Type System 300MB
Partition 2 is Type Unknown 300GB. This is my Linux partition I'm pretty sure.
Partition 3 is Type Primary 175GB.

So the drive is alive and all and I can be spotted in diskpart utility. But is not listed as a block device in Linux. What does that mean and how do I proceed from here? 

 

At least now you know the drive is alive outside of bios.

 

Diskpart is not the best at identifying Linux partitions iirc. So I would start by running some recovery tools from a Linux live media. Try to recover partition table first. Hopefully that is all that is corrupted. Since Manjaro live is being mean, test a different distro if you still can't see the drive there.

 

If you don't have any vital data stored or have backups, you can consider to wipe the drive and start from scratch.

(After sel disk # in diskpart you can type clean. Fair warning: this will wipe the boot sector of your drive without asking for confirmation)

[Limpuls]

34 minutes ago, Pandur said:

At least now you know the drive is alive outside of bios.

 

Diskpart is not the best at identifying Linux partitions iirc. So I would start by running some recovery tools from a Linux live media. Try to recover partition table first. Hopefully that is all that is corrupted. Since Manjaro live is being mean, test a different distro if you still can't see the drive there.

 

If you don't have any vital data stored or have backups, you can consider to wipe the drive and start from scratch.

(After sel disk # in diskpart you can type clean. Fair warning: this will wipe the boot sector of your drive without asking for confirmation)

Okay, thanks a lot for the help. I will continue to investigate. You already helped more than I expected as my original answer was answered long time ago. Thanks!