I run Unraid as my NAS OS and wasn't home during a power outage. The UPS ran out and thus the system lost power. I have three drives in Raid 0 (i know i know ... i will have a mirrored cache from now on) as my write cache: a 1.9TB Seagate SAS SSD and two old 840 Pro in the 512GB variant. One of the Samsungs survived the outage, the other has no sign of life. I've already tried the following:
- power cycle -> leave it on SATA power (no SATA data) over night in hopes that it will self correct any issues, then unplug for an hour to drain all caps, and plug it into power only for another hour or two. no success
- check the BIOS -> no sign here either. my PC doesn't even post when the drive is plugged in. it gets stuck on the Asus splash screen
- pray to the gods of the almighty SSD heaven in hopes that a miracle happens and she springs back to life
- I have checked for shorted caps and dead resistors on the PCB. it's a little tough to distinguish between inductors and resistors as the color is BARELY any different. no burnt or popped components either.

Note that when I plug in the drive the controller gets warm to the touch with the NAND chips getting a little warm after a while. So it seams like power delivery isn't completely dead.

Is there ANYTHING I can do from here? The data is mildly important. It's a mixture of work files but also the fact that I will probably loose all the other data on the other drives as well since I ran the Cache in Raid 0 BTRFS...

Looking for any tips! Thanks in advance.

Given the problem you will be why I use JBOD with hard disks and SSD alike.

From what I gather is that you have some dysfunctional SATA SSD.Power spikes can often destroy SSD storage fast.I never bought many SATA SSD instead I have lots of M.2 SSD which have accumulated. I use thunderbolt M.2 enclosures which make life easier as this allows fast backups.

If you post a photo or 12 of the affected devices somebody can usually point you to fixing them.

Some of my machines have SD card slots soI use them for file history which is able help recover when something dies.

_________________
Orico DS500C3-US-BK 5-disk 3.5" SATA, iTGZ 4x M.2 2280 Thunderbolt enclosure NVMe/SATA, etc. etc.

Thanks for the input!

That's a very good idea with the pictures. I attached the front and back side of the drive (I hope it worked). Maybe someone could point out spots to measure voltages. From what I measured the drive pulls about 0.8 - 1W when powered on.

_________________
A backup a day keeps DR away.

Hey @fzabkar,
I saw that PDF during my research as well. However, the one in the manual is the 840 EVO. I'm running an 840 Pro and the PCBs differ quite a bit. I haven't had the time yet to fully dig through it and maybe discover some important similarities but for the time being I actually managed to rescue some very important data from the other two Raid0 BTFRS drives that were used as an SSD cache. Will give the document you sent me a good read on the weekend tho!
Thanks a lot!

Just an FYI for anyone reading through this who also had a Raid0 BTFRS Cache in their system and one of the two or more drives died:
there's a good chance you can recover parts of your data! As mentioned in my post, I had three drives. One 1.9TB and two 500GB. One of the 500GB Samsung 840 Pros died during a power outage. I thought all was lost since BTFRS stripes data.

HOWEVER. BTFRS does not stripe everything similarly. From what I understand, it's as follows:
-> everything gets split into 64kB chungs
-> if file <1GB, all 64kB chunks remain on one drive
-> if file >1GB, chunks get "zig-zagged" across two
-> if file WAY more than 1GB, chunks get striped across more and more drives

So what does that mean for data recovery? BTFRS prefers the largest drive for storing the meta data tree. If the meta data tree is in tact, you are in pretty good conditions to recover some good amount. Linux has a command in terminal called "BTFRS rescue" that looks for the meta data tree and checks whether it can find all the necessary 64kB chunks on the surviving drives. It can then splice them back together and store them on another drive.
In my case, I was in luck because about 300GB of my files were either stored on the 1.9TB, or on the remaining 500GB, OR they were striped across those two drives. The files I couldn't recover were the ones either stored on the dead drive, striped across the dead and the other 500GB drive, or striped across all three drives.
But that's better than nothing!

I got all my VMs and Dockers back since small files that are only a few kilobytes big are always stored in one and most likely stored on the largest drive to keep everything balanced.

Give the "BTFRS rescue" command a quick google. I am not knowledgeable enough to share code here and not mess something up hahahhha

Sorry, I goofed. You can measure the voltages at the following test points.

_________________
A backup a day keeps DR away.

oh my!!! i am so terribly sorry. i didn't see your reply as it was very soon after my message and then probably didn't send a notification e-mail...
thank you for the help @fzabkar ! it's much appreciated.

so here's what i measured:
- 5V in & out: 4.78V
- V1: 2.93V steary
- V2: 1.8V but occasionally short dip to 1.75V then recovers to 1.8V over the span of about 1 second
--> after measuring everything a second time a minute later it is now stable at 1.8V
- V3: 1.9V steady
- V4: 1.05V steady
- V5: 1.15V steady
- V6: 1.254V steady

i then opened my known good 840 Pro and measured the same spots just to compare it to a working drive. same voltages. just the 5V rail a little higher as it was plugged into a Sata power plug that was closer to the PSU. the rest is identical.
then, i tried to read the PDF you sent me earlier in order to figure out whether something is wrong with the voltages, however, since they are identical to the working drive, i figured those couldn't be the issue and didn't invest too much time into getting into every detail of said PDF.