Jump to content

Fixing broken raid 0?

biotoxin

I want to open by saying I have a full system image from just hours before everything broke and this is more of a learning computer that I somewhat intentionally break almost every month. By learning to fix everything I can help others. 

 

So I was removing a GPU from the system and taking one from crossfire and moving it up. I decided to rearrange some hard drives while I was at it. When I'm done I get a message saying reboot and select device. My first thought was I goofed some cables or boot priority changed, after checking priority and trying all cable order combinations I checked to see if my 3 ssd raid 0 was intact which according to the system it is. 

 

At this point rather than jumping on my normal solution I thought I'd try to fix it proper. I haven't done anything else yet because I want ideas. My first thoughts are break and rebuild the array, if I needed data I'd use software to simulate the array and pull the data, and my third option being clean install and restore but that sucks as a solution. 

 

I'd like to figure out why it broke, and can it be restored better through maybe diskpart for example. 

 

With all the cable play I can't be sure right now which disk is first in the array 7,and i haven't checked yet to see if the array partition is active, so that might be first priority. 

 

 

Spoiler

CPU: TR3960x enermax 360 AIO Mobo: Aorus Master RAM: 128gb ddr4 trident z royal PSU: Seasonic Prime 1300w GPU: 5700xt, 5500xt, rx590 Case: c700p black edition Display: Asus MG279Q ETC: Living the VM life many accessories as needed Storage: My personal cluster is now over 100tb!

Link to comment
Share on other sites

Link to post
Share on other sites

drives are fine, it's just a software error for sure

I'm working in ubuntu live disc for the moment to poke around

looks like just a corrupt gpt table, backup gpt table is fine maybe I can work from that?

 

17 minutes ago, Sierra Fox said:

If your array died...MAAAAYBE, depending if there's even a way to rebuild the same array, a new array would not be able to extract the data across two drives if it doesn't know where to look. if a drive died, absolutely not

if you break an array and quick initialize the EXACT same array it tends to fix such errors, but since I already know how to and have done so before I'm looking for a new way, maybe a better way

Spoiler

CPU: TR3960x enermax 360 AIO Mobo: Aorus Master RAM: 128gb ddr4 trident z royal PSU: Seasonic Prime 1300w GPU: 5700xt, 5500xt, rx590 Case: c700p black edition Display: Asus MG279Q ETC: Living the VM life many accessories as needed Storage: My personal cluster is now over 100tb!

Link to comment
Share on other sites

Link to post
Share on other sites

31 minutes ago, TrigrH said:

did you leave your sata mode in RAID the entire time?

 

or did you set it to raid, make the array, then back to AHCI?

 

EDIT? are u on an AMD system?

yes AMD system, yes switched from raid to ahci for the last year-ish

switched back to raid for a minute to verify array wasn't completely broken, now working in ahci again for the moment

 

edit: also these are only connected drives on chipset, all other drives connected from a separate raid controller currently unplugged while I'm messing around. the sabertooth 990fx chipset has a bad habbit of dropping drives in an array if you connect other drives not in that same array on chipset

Spoiler

CPU: TR3960x enermax 360 AIO Mobo: Aorus Master RAM: 128gb ddr4 trident z royal PSU: Seasonic Prime 1300w GPU: 5700xt, 5500xt, rx590 Case: c700p black edition Display: Asus MG279Q ETC: Living the VM life many accessories as needed Storage: My personal cluster is now over 100tb!

Link to comment
Share on other sites

Link to post
Share on other sites

Just now, biotoxin said:

yes AMD system, yes switched from raid to ahci for the last year-ish

switched back to raid for a minute to verify array wasn't completely broken, now working in ahci again for the moment

thats why it isn't working (cos you used AHCI), you can try rebuilding the raid under the exact same settings, your data shouldn't be lost.

 

Otherwise RIP

Link to comment
Share on other sites

Link to post
Share on other sites

2 minutes ago, TrigrH said:

thats why it isn't working (cos you used AHCI), you can try rebuilding the raid under the exact same settings, your data shouldn't be lost.

 

Otherwise RIP

first up I need to confirm what the proper drive order is, again I played with cables thinking I'd accidentally done something there, then I could do that

Spoiler

CPU: TR3960x enermax 360 AIO Mobo: Aorus Master RAM: 128gb ddr4 trident z royal PSU: Seasonic Prime 1300w GPU: 5700xt, 5500xt, rx590 Case: c700p black edition Display: Asus MG279Q ETC: Living the VM life many accessories as needed Storage: My personal cluster is now over 100tb!

Link to comment
Share on other sites

Link to post
Share on other sites

Just now, biotoxin said:

first up I need to confirm what the proper drive order is, again I played with cables thinking I'd accidentally done something there, then I could do that

if you didn't take a photo of your original order and stripe = silly

not using a UPS = silly

ahci = stupid, always stay in RAID for a RAID.

 

also you don't get TRIM support on AMD motherboards, only intel.

Link to comment
Share on other sites

Link to post
Share on other sites

Just now, TrigrH said:

if you didn't take a photo of your original order and stripe = silly

not using a UPS = silly

ahci = stupid, always stay in RAID for a RAID.

 

also you don't get TRIM support on AMD motherboards, only intel.

I do have the details of stripe, hard to tell which drive is which thought

spent good money on my ups, though irrelevant to current

I'll look more into ahci situation, my understanding is it shouldn't be a negative impact?

and trim support is on both now just not over raid, my understanding though is that maybe intel does trim over raid? I'd have to verify

either way not much of a terrible concern for me as by the time the drives die I'll be well onto better systems working with things like nvme

 

Spoiler

CPU: TR3960x enermax 360 AIO Mobo: Aorus Master RAM: 128gb ddr4 trident z royal PSU: Seasonic Prime 1300w GPU: 5700xt, 5500xt, rx590 Case: c700p black edition Display: Asus MG279Q ETC: Living the VM life many accessories as needed Storage: My personal cluster is now over 100tb!

Link to comment
Share on other sites

Link to post
Share on other sites

Just now, biotoxin said:

I do have the details of stripe, hard to tell which drive is which thought

spent good money on my ups, though irrelevant to current

I'll look more into ahci situation, my understanding is it shouldn't be a negative impact?

and trim support is on both now just not over raid, my understanding though is that maybe intel does trim over raid? I'd have to verify

either way not much of a terrible concern for me as by the time the drives die I'll be well onto better systems working with things like nvme

 

The serial number is on each drive.

 

Intel passes the trim command through raid 0 on 7 series motherboards and newer (Z77 etc)

 

When you are running in RAID mode, there is significantly less chance of shit like this happening, hell i've moved to entire new system and the motherboard has found my raid.

Link to comment
Share on other sites

Link to post
Share on other sites

Ok well this thread has gotten out of hand....  If you have changed the order of the cables then yes this will be the cause of your broken array. (Mobo expect Drive 1 to be in SATA port X) etc. You can attempt to fix it by trying to get the correct order again. What does Intel RST say about your RAID?

 

To check if your RAID0 is trimming within windows use this command from a CMD:

fsutil behavior query DisableDeleteNotify

You are wanting the result to = 0, if you get 1 then you DONT have trim enabled.

Intel I9-9900k (5Ghz) Asus ROG Maximus XI Formula | Corsair Vengeance 16GB DDR4-4133mhz | ASUS ROG Strix 2080Ti | EVGA Supernova G2 1050w 80+Gold | Samsung 950 Pro M.2 (512GB) + (1TB) | Full EK custom water loop |IN-WIN S-Frame (No. 263/500)

Link to comment
Share on other sites

Link to post
Share on other sites

heading to work, when I'm back I'll update with pics of exactly what I'm talking about, for the sake of potential benefits

 

Edit, I've got an extra hour, here's my album so far

https://goo.gl/photos/N2ffCFMX8hXCD4jN8

Spoiler

CPU: TR3960x enermax 360 AIO Mobo: Aorus Master RAM: 128gb ddr4 trident z royal PSU: Seasonic Prime 1300w GPU: 5700xt, 5500xt, rx590 Case: c700p black edition Display: Asus MG279Q ETC: Living the VM life many accessories as needed Storage: My personal cluster is now over 100tb!

Link to comment
Share on other sites

Link to post
Share on other sites

Moved drives around again, think I've got them in order now, no real clue though, looking up ways to manage partition tables on raid, and how to verify raid drive order

Spoiler

CPU: TR3960x enermax 360 AIO Mobo: Aorus Master RAM: 128gb ddr4 trident z royal PSU: Seasonic Prime 1300w GPU: 5700xt, 5500xt, rx590 Case: c700p black edition Display: Asus MG279Q ETC: Living the VM life many accessories as needed Storage: My personal cluster is now over 100tb!

Link to comment
Share on other sites

Link to post
Share on other sites

Updated pics, can now see array in Windows Recovery with data, integrity not verified 

 

Can't fix startup yet either and now it actually is time for work

But I'm feeling hopeful 

Spoiler

CPU: TR3960x enermax 360 AIO Mobo: Aorus Master RAM: 128gb ddr4 trident z royal PSU: Seasonic Prime 1300w GPU: 5700xt, 5500xt, rx590 Case: c700p black edition Display: Asus MG279Q ETC: Living the VM life many accessories as needed Storage: My personal cluster is now over 100tb!

Link to comment
Share on other sites

Link to post
Share on other sites

Bootrec doesn't see the OS, diskpart does but doesn't seem to help me fix anything, the partition table looks OK, and I've rebuilt the bcd boot all to no avail, it's like 99% OK out just won't actually boot into the OS 

Spoiler

CPU: TR3960x enermax 360 AIO Mobo: Aorus Master RAM: 128gb ddr4 trident z royal PSU: Seasonic Prime 1300w GPU: 5700xt, 5500xt, rx590 Case: c700p black edition Display: Asus MG279Q ETC: Living the VM life many accessories as needed Storage: My personal cluster is now over 100tb!

Link to comment
Share on other sites

Link to post
Share on other sites

I'm still fussing around 

Moved back to ahci to experiment and noticed how the drives disappeared so I do a disk list and notice they're invalid 

So I updated the Google album again 

 

This week was hectic, maybe now I can focus on fixing 

Spoiler

CPU: TR3960x enermax 360 AIO Mobo: Aorus Master RAM: 128gb ddr4 trident z royal PSU: Seasonic Prime 1300w GPU: 5700xt, 5500xt, rx590 Case: c700p black edition Display: Asus MG279Q ETC: Living the VM life many accessories as needed Storage: My personal cluster is now over 100tb!

Link to comment
Share on other sites

Link to post
Share on other sites

Apparently the answer as always is testdisk, it found something, album updated 

 

Spoiler

CPU: TR3960x enermax 360 AIO Mobo: Aorus Master RAM: 128gb ddr4 trident z royal PSU: Seasonic Prime 1300w GPU: 5700xt, 5500xt, rx590 Case: c700p black edition Display: Asus MG279Q ETC: Living the VM life many accessories as needed Storage: My personal cluster is now over 100tb!

Link to comment
Share on other sites

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×