RAID 0 SSD unresponsive under high load

[FCM]

Hello,

 

I've been running a RAID 0 SSD using Intel RST driver. Under high sustained transfer (eg: downloading a game via steam), the app becomes unresponsive and windows event log throws multiple "reset to device, \device\raidport1, was issued."

 

I generally issue a hard reset here, to recover the system.

 

I've changed all the sata cables, but the issue remained.

I've also disabled Link state Management in the power profile.

It could be tied to write back cache being enabled, I'm gonna play around with those settings to see if anything changes.

 

In the end, I decided to forego the raid 0 setup, switched to AHCI and the used Windows 10 spanned volume feature to create a volume split across the SSDs. I expect some performance drop, given that's basically software raid.

 

SSDs are new, barely 1-2 month old and have been used under normal conditions on a gaming machine.

 

 

The issue however disappeared, I've deleted several games from the Steam library and re-downloaded them just to be sure.

 

Right now, I'm suspecting two things:

- a software bug in Intel RST driver, or Intel Z370 Raid firmware or SSD firmware.

- One of the SSD is dying.

 

Any thoughts?

 

I've searched a bit on the web and it seems that one approach would be to actually let the "reset device" time out, at which point Intel RST would indicate if and what SSD was removed from the RAID set-up. That would indicate the faulty drive.

 

I would do that, but first I want to see if there are other options which wouldn't require data wipes.

 

Specs:

MSI Z370 + 8700K

2 x 850 EVO 250GB.

[dalekphalm]

You should check the SMART data on both SSDs and run diagnostics. 

[FCM]

Any specific value or diagnostic tool? Samsung magician says they're ok.

[dalekphalm]

On 2/14/2018 at 7:45 PM, dalekphalm said:

You should check the SMART data on both SSDs and run diagnostics. 

 

6 hours ago, FCM said:

Any specific value or diagnostic tool? Samsung magician says they're ok.

Please see above. Check the SMART data on both drives (check for URE's, other errors, reallocated sectors, etc).

 

Then, run a SMART Short, and SMART Long test on both drives.

 

With Samsung Magician, did it do a full drive scan? Even on an SSD this should take at least a few minutes if not longer.

 

If all tests come back okay, then the drives themselves are unlikely to be the problem.

[FCM]

This issue went away once I've updated my bios to the latest version (Z370 Krait Gaming).

 

That bios updated specifically called out an updated from RST 15.9 to 16.

 

So it was most likely a raid irmware bug.