Toshiba PX05SRB SSD Failures

[nmap]

Hi all, just posting to see if anyone has any other possible solutions / possibly to raise awareness of these drives failing.


I have a several Toshiba PX05SRB SSD’s that are all (except one) reporting to a Dell RAID card that they’ve failed, and when they are plugged into a SuperMicro chassis with a different SAS controller, the drives report a capacity of 0 Bytes.


Screenshot from the SuperMicro SAS Controller:


Screenshot of the email from iDRAC:


I’ve tried:

• Inserting the drive in a different slot - makes no difference
• Using the SuperMicro’s SAS controller to verify or format the drives - it instantly fails
• Booting from an Ubuntu Live ISO to work on the drives - all options for disk management were greyed out

I do see that IBM says there is a required firmware update these drives to prevent premature failure, but I don’t have one of their arrays to apply said update.

Required drive firmware upgrade for 1.92 TB and 3.84 TB Tier 1 Flash drives with 6402 firmware

Specific 1.92 TB and 3.84 TB Tier 1 Flash drives, running 6402 firmware, are exposed to an issue that can cause drives to fail prematurely - potentially on multiple drives in an array, causing loss of data. Upgrade to the latest 6406 drive firmware is strongly recommended, to prevent this issue.

www.ibm.com

I’m guessing all of the drives are toast, but I’d really like that to not be the case haha


Thanks!

 

[jamtech]

I was researching before purchasing a couple of used PX05SRB for one of my Supermicro boxes and found this thread - glad I did!
Think I'll give them a miss.

 

[nmap]

If I could get ahold of a firmware update, they'd probably be pretty good lol
Unfortunately, I still haven't found one, and Kioxia support said they couldn't give me one.

 

[CyklonDX]

by the looks this typically happens when resistor/capacitor fails and you get wrong voltage coming in or one of the memory chip failed. I've seen it on all kinds of ssd's. (it always came from servers with issues - so likely problems with power distribution/mobo causing all kinds of other issues)

I have over 100 of those - exact models at work with dell servers, and we only had 1 failure past last 5 years (at least 48 disks are heavily used for zfs in r730s).

 

[Whaaat]

firmware 0104 for those who care

 

[nev_neo]

@nmap Did the firmware update work ?

 

[drdepasquale]

I've experienced three failures of these particular drives within 30 days of purchasing them. I had these SSDs connected to an LSI 3008 controller as jbod. Fortunately, the eBay seller accepted my return.

 

[nev_neo]

I recently purchased a couple of the netapp x357 variants of the Toshiba drives. They came with the NA54 firmware already loaded.
Does anyone know if that is the latest firmware for these drives ?
The Samsung version which i also have - I had to update those to NA54 from NA52 that they came with to save it from the 40K bug.
Does anyone have any insight on this ?

 

[wmarek]

Would be curious if anyone found a solution to jumpstart these drives.

I have 4 servers pulled from production. they ran for roughly 4 years and have been offline for 3. Not fully powered off but left in the rack for plugged in. When powered down I think remaining write endurance 60%.

I wanted to spin up a test cluster today so powered up the servers, but it was taking forever for them to boot.
Boot drives 2 x 800GB Toshiba PX05SMB080Y
Data drives 6 x 1.92TB Toshiba PX05SRB192Y

All 8 boot drives show as healthy and happy
5 out of the 24 1.92TB drives are healthy. So 19 drives are showing failed with a size of 0KB

Manufacturing year and week are the same for all of the drives.
Moral of the story is don't turn them off?

 

[CyklonDX]

take one apart and check for shorts, see which component is dead. Sometimes capacitors can act like resistors after reaching certain temperatures with no electricity moving through them - but then moving enough electricity changes them back into capacitors.

 

[wmarek]

I did take one apart, tested the capacitors I could look at. they were showing what I would expect them to be showing.
Since they had been powered down, they had been kept at room temperature the entire time, stored in the IT closet of the business.
When connected, the drives show all the normal information like SN, model number, firmware version, all that, but show failed, 0KB size.
Edit: It also happily took a firmware update. just crazy to have 80% of the drives be in a similar state.

 

[CyklonDX]

did you find any short? I mean this is really odd.

 

[wmarek]

No shorts that I could identify, but what would cause shorts across 4 servers and 19 out of 24 drives? systems were properly shut down, un-cabled, sat in a rack in a closet the rest of the time. It is really odd, like a manufacturing defect or something? not like Toshiba/Dell will do anything about it out of warranty though.

 

[CyklonDX]

would you be able to send me 2 of the dead ones? (private msg me)

 

[Charles S.]

I just found this thread and noticed the same thing. Servers that have been running these 960GB PX05SRB096Y drives with no issues for years all of a sudden have 10-20 drives across a few node vSAN cluster where they show 0 bytes. This just happened almost overnight with firmware version AS0B. Upgraded to AS10 which is the last release of the firmware from 2022 but still showing 0 bytes. It would be great if anyone has experience with how to recover the capacity if at all possible. No data recovery is needed, just the use of the drives again. I heard of this happening with Intel drives in 2010 generation and they provided recovery instructions through their own tools, but I am not seeing anything from Toshiba. Thanks.