Raid 1 nvmes / x570

[Cramig88]

Hello all, my pc buddy isn’t responding. Probably out doing stuff, oh well.

Moving on..

 

what happened?  
Was using win 10 pro, had WSL kali, and Ubuntu going, both upgrading/fixing dependencies, and dist-upgrade. Had 2-3 tabs open in Firefox, one for college (blackboard), one was another for a radio station for music/background noise, and another prob related to college stuff.  Word was also open so I was probably doing a virtual lab on JBLEARNING.com. So, follow instructions, do a screenshot and paste into word. 
As all of this was going on, hey my amd 6700xt and 6600 have a new driver available, let’s install it!  So it’s installing while all this is going on, then BSOD, not sure how far into the install but I specifically remember the error code was amd*(gpu-related).dll, yada yada please wait while we report to msft and reboot for you.

Pc reboots, no gpu output as expected. The ROG logo showed up (booting windows), and screen goes blank. I presume Windows loaded fine and was on the login screen, but of course couldn’t see it.

fudge right?  No onboard output, so wtf.
 

what have I tried?

 

-Reboot, then I force it to do the Windows “diagnose” BS. Go into CMD from there as the objective is clearly to remove the AMD drivers and fall back to basic display adapter.  Well, diskpart shows my two nvmes when in reality the two 1tbs should be “one” physical drive as it’s raid 1. Hmm, looks like this pre-Windows boot isn’t loading the NVMe RAID drivers so C:\ isn’t accessible. Damn!  Can’t even go back to a restore point!

-shoved a bootable win10 usb install disk into pc and boop, fires up. Goto the part where it wants u to pick a drive to install on. Hey there’s my 2 nvmes!  Then I shove in my AMD RAID usb, install the drivers, bam, there’s my partitions. Cool. Shift+F10 (I think) to get into CMD.  Well fudge, the raid drive is there just not “mounted”, as in can’t access the media/file system. Shoot!

-rotate my chair 90 degrees, now I’m playing off my server (2022 datacenter) off a kvm switch as it’s in the basement not the 2nd floor office. Hmm…. Ok let’s make a Windows 2 go flash drive. Fire up Rufus, make the drive, shove into pc n reboot.  (Long time later) windows desktop. Ok, now we’re getting somewhere?  Diskpart shows the two nvmes yet again. As expected, no access. That’s ok, I’ll just install the raid chipset drivers and the raidx utility from AMD. Reboot, wtf?  Still can’t access the “drive”.

 

-Alright so now obv we have an issue. Shutdown, insert my hiren boot “cd”, Windows PE fires up fine. Of course I can’t install the NVMe raid drivers since it’s such a lightweight version of Windows (figured I’d try). Sure enough diskpart shows the nvmes again but no access to C:\ or whatever letter it would assign if the array was functioning properly.  Now I’m using some tool for data recovery to “deep scan” one of the drives. Hopefully I can just get to my desktop folder and copy my “college” folder as that’s where all my assignments and stuff are. That’s really all I care about salvaging. Probably less than 200 meg of data.  It’s still scanning, drive is in good Health and it’s happy. (Still going as of now)

 

-the weird thing

i have 8x 1tb Samsung SSDs, 7 are EVO’s and 1 is a QVO - all identical firmware. Anywho, 4 of those are on raid 10 which pull up as a drive letter (I use it to store backups as for some reason the SATA raid never has a problem, even without drivers windows 2 go can see it and I can access). Also the other 4 drives are deticated hot spares (global spares?) just in case - right??  Been there before, had a SSD/HDD setup before on 10 and one drive pooped, another, then u guessed it, everything went to the trash and I swapped my config.

 

-what I probably should have tried

remember initially how windows WOULD boot but no output?  Assumed it was sitting at the login screen?  Well, should have tried to RDP from the server to the pc. It has a static IP of 192.168.1.19.  Maybe I could have pulled up the GUI off the server and reinstalled the AMD video drivers

 

-here is also a consideration, but I’d like solid advice on this one as again, it’s crucial data I want to recover - a whopping 200mb or less.

go into bios, set to ACHI (on nvme’s), make one of two dynamic so I can *maybe* access the drive. I mean, they’re raid 1 so it’s a 100% mirror right?  In theory it should work

ALTHOUGH

when setting up the array in the bios (NVMe, sata, don’t matter) the drives need to be unformatted Blah blah. Also, they need to be “initialized” in preparation for making the array.  This leads me to beleive the bios is writing data to the drives making them say, hey I’m gonna be a member of an array, I’m special. Then ya make the array and now they’re “bound”, blood brothers I guess.  So with that said, the drives must have a portion of data written saying “that other guy who matches my data signature is also me, I can’t work alone”. Of course whatever data it is, has to be stored at the very beginning of the drive in raw storage. Idk let’s say it’s only 128kb worth of code or something. U have the NVMe firmware and whatever else specific to the drive (serial number, sequence data blah blah) then that part of code associating this drive to that one, and then of course all your data - efi partition, NTFS partition, and all the other stupid ones Windows tosses in there.

 

-so where to go from here?

ideally, I’d like to break the array. I’ve contemplated simply removing one of the sticks (after disabling raid and selecting ACHI) and see if it can figure out wtf is going on. Or perhaps leaving raid on (that little part of data right?), then removing one of them. Maybe bios will see one drive is “down” and figure it out?

 

-things to consider

in bios AND windows 2 go, the NVMe raid array is “offline” - zero clue why. I didn’t mess with anything in there as of yet. Well actually no, I did clear the cmos but that still retained the raid arrays. Again, the sata array shows up totally fine, it’s online and happy, all the spares show up too. But wtf is going on with the NVMe raid??  NOTE, the cmos was not cleared until after I did the whole win2go/install raid drivers/reboot/check to see if I can access the “drive”

 

any ideas?  It be nice if I could just salvage my college folder at a bare minimum. After that, then try to recover one of the disks as a whole (so Windows can physically mount it as a letter) and remove that half-installed driver.

also under disk management (computer management, whatever), the drives do pull up but for me to assign a drive letter I need to convert one of em to dynamic, from gpt. From what I’ve read online when u go dynamic u can’t convert back to gpt. Basically, I’m not quite sure if it’ll help or set me back even further.

after all, it would be nice just to get windows back n working like it was yesterday prior to the BSOD. I don’t mind a fresh format either, so long as I have my college folder (finals are in like a week).

 

-Pc specs

x570 asus/rog dark hero WIFI on 4402 bios (latest last I checked, feb 02 of 2023 I beleive)

5900x OCd to 4.25

bclock set to 101.25 (could be lower now, probably just 100, not that it matters)

4x Corsair dominators, 3600, 64gb total (16g each) with docp enabled

Xfx 6700ti and another xfx 6600

1000w seasonic 80 gold psu

a (bleep) load of watercooling stuff. Let’s just say over 1800mm worth of radiators.  Yes my case is packed!

a crapload of fans, forget the brand, but non RGB. I do know they do 2500rpm, the 120mm’s at least. Same model line, 5x 140mm fans, just a tad slower but fastest fans I could find. They all match so yeah, irrelevant but good to know?

2x Kingston fury (rampage?) 1tb nvmes, latest firmware last I checked.  Made in same country, same model, everything is identical about those aside from serial number

8x Samsung’s on raid10 (sata) as mentioned, no need for further details on those. They all match but one which is one of 4 global spares

 

hardware NOT installed yet, or ever (could be helpful for anybody that has ideas)

-one gigabyte NVMe pci-e 4.0 16x enclosure packed with 2x 980 pros and 2x Kingston rampages (fury’s, whatever)

here’s the copy/paste off Amazon if you’d like to look it up

”GC-4XM2G4 (AORUS Gen4 AIC Adaptor, Full PCIe 4.0, Advanced Thermal Solution for PCIe 4.0 SSD)”

^again, never installed in pc, never setup, nothing. Just put the drives in the enclosure, closed it up and it’s sitting on top my case waiting to be installed after this mess is figured out.

oh and should this get fixed, should I just put all 4 of my Kingston’s into that and put the two 980s on the motherboard?  I plan on raid’ing the enclosure thingy - IF it has an onboard raid controller (that will eliminate the need for drivers as it’s made for my motherboard)

Should it work, assuming butrificatuon isn’t a problem, put my main OS on that?  How should I configure my HD’s in relation to my default boot drive?

 

geez sorry for the long story. Still waiting on this hiren app to deep scan the drive

anyway, thanks in advance. I really really need some help on this one, very very desperate

 

OH and the server is literally a server, idk if it helps with ideas but yeah. It’s got like 10 NICs on her, U2 form factor, dual Xeon somethings, 384ish gb of ddr3, registered, ecc, blah blah. There’s also a quadro rtx 4000 in it, which I probably should have just shoved in my pc had I thought about it back then…. Full ipmi access, latest firmware. Supermicro something. So I could always do pxe boot, although I don’t quite understand how to mount a iso thru the interface as it wants a location, like \\yadayada\this.iso. I imagine it wants either \\Server\SAS-RAID\Downloads\ISOs\windows10.iso OR like linux, \\pci{03840283957302947}\somestuff\morestuff\noidea\modelnumber\sda1\Downloads\ISOs\windows10.iso

^its also raided, Intel 120gb SSD as boot, 6x 1tb sata3 SSDs on raid 0, also another array, 8x seagate somethings, 2tb each on raid 0, SAS drives

plus a WD direct access storage, 14tb strictly used ONLY for backups. I mean we got a lot of terabytes here on raid 0, so B/Us are done frequently just in case

For right now it’s just a storage server, static ips are set. 2022 datacenter. After college is done with it’ll become a lot more, and I’ll toss in my other 2 laptops as servers doing different roles. (That’s why I have the kvm balun, once the laptops are setup as servers they’ll be wired to the supermicro and I’ll just RDP to each of them as needed)

 

ugh ok that’s my bit. Help!!