Rapid Storage Technology
Intel® RST, RAID
2047 Discussions

iRST driver (17.7.0.1006) unstable ... BSOD caused by iaStorAC.sys?

dmuso1
Beginner
4,942 Views

This latest iRST driver (17.7.0.1006) seems unstable with my RAID 0 configuration. When I run disk repair utilities on my SSD array such as HD Sentinel PRO and Chkdsk, I get the BSOD about half-way through the scanning process! It says "iaStorAC.sys" failed with a stopcode of SYSTEM_THREAD_EXCEPTION_NOT_HANDLED ( see attachment). I use two (2) 500GB Samsung 950 PRO SSDs to create a 1TB RAID 0 array.

 

I have no problems with the default windows iRST version 15.44.xxx but its slow compared to version 17.xx. What should I do to fix the BSOD issue?

 

PS

I'm not sure if this is related but I started to notice the BSOD after I performed a FW upgrade on my SSDs the day before. Not sure if these two events are connected but I doubt it since the Samsung SSD FW has been out for a couple of years.

0 Kudos
13 Replies
Leonardo_C_Intel
Moderator
4,409 Views

Hello dmuso1

 

Thank you for posting in the Intel Community.

 

I would like to have more information about your system configuration. please provide me with the following reports:

 

·        Provide the Intel® System Support Utility for Windows* (https://downloadcenter.intel.com/download/25293/Intel-System-Support-Utility-for-Windows-) report.

·        Provide the System Reports for Intel® Rapid Storage Technology (https://www.intel.com/content/www/us/en/support/articles/000006351/technologies.html)

 

NOTE: completing the firmware update on the SSD has to be done without having the disk on a RAID structure since you mentioned that you have a RAID 0 you will have to back up the data, destroy the RAID, complete the firmware update then rebuild the RAID and copy the data back.

 

Regards,

Leonardo C.

 

Intel Customer Support Technician

A Contingent Worker at Intel

0 Kudos
dmuso1
Beginner
4,409 Views

Thank you Leo for responding and I've attached the files you've requested...

 

First of all, I reverted to the previous driver, version 17.5.2.1041 and everything works fine as I had expected. Disk utilities work just fine with no issues and no more BSOD issues. I checked the release date of the latest driver 17.7.1.1006(?) and it was Dec 28th, 2019. It seems (to me) that it was a rush job, not completely tested but had to be released before years end, for some reason? When I reverted to the 17.5 version, I was informed that I would be replacing a newer files (iaStorAC.sys/ini/.xxx) with an older one for the 17.5 version. It was the only file that generated that message, so it seems all other files between 17.5 and 17.7 are identical, except those that belong to iaStorAC.xxx drivers.

 

Secondly, when I updated the SSD firmware, I disabled the raid array in BIOS and used the AHCI mode. I had a bootable thumb drive with the updated firmware and the update was done in two steps, one for each 500GB SSD drive. Afterward, I re-enabled the raid array and booted normally and all was ok. No more BSED issues when I run various drive scan and repair utilities. With version 17.7, they all crashed with the BSOD around the 50% - 52% scan/repair point, right where I had several bad and unreadable drive sectors. Now, all are repaired with iRST version 17.5 . I would like to install 17.7 once its fixed, but these public FW releases just before year-end are suspect and not advisable.

 

Thirdly, I'm only able to attach 1 file at a time(?).

 

Thank you for your time!

 

Regards,

David Musoke

 

iRST_System Report below:

 

System Report

System Information

OS name: Microsoft Windows 10 Pro

OS version: 10.0.18363

System name: DESKTOP-0S95Q5J

System manufacturer: Micro-Star International Co., Ltd.

System model: WT72 6QL

Processor: GenuineIntel Intel64 Family 6 Model 94 Stepping 3 2.601 GHz

BIOS: American Megatrends Inc., E1782IWS.121

PCH: 0xA150

 

Intel® Rapid Storage Technology enterprise Information

Kit installed: 17.5.2.1041

User interface version: 17.5.2.1041

Language: English (United States)

RAID option ROM version: 14.6.0.2285

Driver version: 17.5.2.1024

ISDI version: 17.5.2.1041

 

Storage System Information

RAID Configuration

 

Array Name: SATA_Array_0000

Size: 976,772 MB

Available space: 3 MB

Number of volumes: 1

Volume member: Intel SSD 950PRO

Number of array disks: 2

Array disk: 0025_3853_61B0_1A3E.

Array disk: 0025_3853_61B0_1A25.

Disk data cache: Enabled

 

Volume name: Intel SSD 950PRO

Status: Normal

Type: RAID 0

Size: 976,768 MB

System volume: Yes 

Data stripe size: 16 KB

Write-back cache: Off

Parity errors: 0

Blocks with media errors: 0

Physical sector size: 512 Bytes

Logical sector size: 512 Bytes

 

Hardware Information

 

Controller name: Intel(R) Chipset SATA/PCIe RST Premium Controller \\Scsi0

Type: SATA

Mode: RAID

Number of SATA ports: 6

Number of volumes: 1

Volume: Intel SSD 950PRO

Number of spares: 0

Number of available disks: 1

Rebuild on Hot Plug: Enabled

Manufacturer: Intel Corporation

Model number: 0x2822

Product revision: 49

Direct attached disk: W77232PX

Direct attached disk: 0025_3853_61B0_1A3E.

Direct attached disk: 0025_3853_61B0_1A25.

 

Disk on Controller 0, Port 3

Status: Normal

Type: SATA disk

Location type: Internal

Usage: Available

Size: 932 GB

System disk: No

Disk data cache: Enabled

Command queuing: NCQ

Transfer rate: 6 Gb/s

Model: ST1000LM014-1EJ164                    

Serial number: W77232PX

SCSI device ID: 0

Firmware: SM36   

Physical sector size: 4096 Bytes

Logical sector size: 512 Bytes

 

Disk on Controller 1, Port 0

Status: Normal

Type: PCIe SSD

Location type: Internal

Usage: Array disk

Size: 477 GB

System disk: No

Port interface: NVMe

PCIe link speed: 4000 MB/s

PCIe link width: x4

Model: Samsung SSD 950 PRO 512GB              

Serial number: 0025_3853_61B0_1A3E.

SCSI device ID: 1

Firmware: 2B0QBXX7

Physical sector size: 512 Bytes

Logical sector size: 512 Bytes

 

Disk on Controller 3, Port 0

Status: Normal

Type: PCIe SSD

Location type: Internal

Usage: Array disk

Size: 477 GB

System disk: No

Port interface: NVMe

PCIe link speed: 4000 MB/s

PCIe link width: x4

Model: Samsung SSD 950 PRO 512GB              

Serial number: 0025_3853_61B0_1A25.

SCSI device ID: 3

Firmware: 2B0QBXX7

Physical sector size: 512 Bytes

Logical sector size: 512 Bytes

 

ATAPI device on Controller 0, Port 2

Location type: Internal

Transfer rate: 1.5 Gb/s

Model: HL-DT-ST BD-RE BU20N                  

Serial number: MD8FALA2147

Firmware: 1.00   

 

Empty port

Port: 0

Port location: Internal

 

Empty port

Port: 1

Port location: Internal

 

0 Kudos
Leonardo_C_Intel
Moderator
4,409 Views

Hello dmuso1

 

I was checking the details that you have provided, fist I would like to thank you for the clarification of the steps that you have taken to complete the firmware update on the SSD, in addition, based on your system details/specifications you are using a chipset from the Intel® C230 Series Chipsets that is a server intended chipset on an MSI* workstation we recommend referring the issue with MSI* support If it is needed MSI will work the behavior directly with Intel.

 

Regards,

Leonardo C.

 

Intel Customer Support Technician

A Contingent Worker at Intel

0 Kudos
dmuso1
Beginner
4,409 Views

Hi Leo … thanks for responding

 

Now to your statement about using a server-based chipset in a laptop such as mine …. First of all, it would be virtually impossible for such a chipset to run on my laptop due to the chipsets extreme power requirements that do not fit within the laptops power budget. MSI has been designing millions of top-notch laptops with the C236 chipset since its first introduction in Q1 2015. The C236 is a mobile chipset for the Skylake family of processors, which my laptop has.

 

Secondly, I believe there's been a mix-up regarding chipsets. The C230 series family of chipsets you alluded to earlier is not a storage controller chipset that handles RAID operations. RAID tasks belong to the Intel SATA/PCIe RST Premium Controller chipset, which uses the buggy firmware in question. According to Device Manager, the C230 series uses completely different driver firmware (10.1.1.45) than the iRST Premium Controller chipset for raid that uses the buggy firmware in question (17.7.0.1006).

 

I hope that clears up the confusion about chipsets used in my system. I'd still want a permanent solution to my original problem. As per the BSOD, the problem lies in the latest version of the iaStorAC.xxx driver.

 

Best Regards,

David Musoke

 

0 Kudos
Leonardo_C_Intel
Moderator
4,409 Views

Hello dmuso1

 

Thank you for waiting.

 

Did you download Intel® RST for Intel® download center side? (Reference link: https://downloadcenter.intel.com/product/55005/Intel-Rapid-Storage-Technology-Intel-RST-)

Can you provide us with the BSOD dumb flog file? This link might help: https://social.technet.microsoft.com/Forums/windows/en-US/684d77b8-0257-426e-a363-54fa54c64a3e/location-of-blue-screen-dump-file?forum=w7itproperf

 

Please be aware that the content on that site is not controlled by Intel®. This information is offered for your convenience and should not be viewed as an endorsement by Intel® for the merchants or services offered there.

 

Regards,

Leonardo C.

 

Intel Customer Support Technician

A Contingent Worker at Intel

0 Kudos
dmuso1
Beginner
4,409 Views

Thanks for the response Leo:

 

A) Yes, I download all my iRST executables from the Intel site you linked to. I cannot afford to get them elsewhere. Also, I use the Intel Driver Support & Assistant utility tool which boots up on windows startup to search for new drivers daily and download them accordingly.

 

B) As for the windows dumb files, none existed on my drive since I had set the windows pagefile size to zero. So all dumb files were stored in volatile RAM which erased them on bootup. I re-installed the buggy version 17.7 and I set the pagefile setting to Auto, so that windows manages my pagefile. I rerun chkdsk /f /r for bad sector repair, if it found any. There was no BSOD seen this time around, which was confirmed with other disk repair utilities I used. I had no bad sectors as they were repaired under the earlier version 17.5 driver, which I've now been running for a about a week.

 

I repeated the chkdsk test again, this time with the Pagefile setting back to zero, as I had it originally when I observed the BSOD errors with the 17.7 driver and re-run all my disk repair tests. No BSOD screens observed, which was a relief. So, I came to the following conclusion.

 

  1. The Pagefile setting (none or Auto) has nothing to do with the observed BSOD issues.
  2. The 17.7 definitely driver fails if there bad sectors on the drive that are in process of being repaired as observed before. The problem is in the iaStorAC.xxx driver(s) as indicated by the original BSOD graphic.

 

As you may imagine, I'm not about to cripple my main SSD drive to simulate a bad sector and attempt a repair under version 17.7. I hope you understand.

 

Regards,

David

0 Kudos
Leonardo_C_Intel
Moderator
4,409 Views

Hello dmuso1

 

Thank you for the response let me look into the details that you have shared. I will be posting back as soon as possible.

Regards,

Leonardo C.

 

Intel Customer Support Technician

A Contingent Worker at Intel

0 Kudos
Leonardo_C_Intel
Moderator
4,409 Views

Hello dmuso1

 

Thank you for waiting.

 

This behavior has been forwarded to engineering and tests are being conducted and any issues will be ironed out in the next release in the meantime:

 

·        Can you roll back the Samsung SSD FW and see if the issue still occurs?

·        Also, was it working fine with before updating the SSD FW on 17.7 drivers?

 

Regards,

Leonardo C.

 

Intel Customer Support Technician

A Contingent Worker at Intel

0 Kudos
dmuso1
Beginner
4,409 Views

Hi Leo:

 

  1. I cannot roll back to the old firmware, since it no longer exists on Samsun's site. This FW was released in mid-2015 or early 2016. I was just lazy and didn't upgrade my NVM. If it was buggy, there'd be lots of bug reports online with this BSOD issue. Its not difficult to emulate if you had a faulty NVM with bad sectors... and Samsung would receive lots of grief over it. Its new FW has been stable for over 4 years.
  2. I don't really know as I upgraded the NVMs first then the 17.7 drivers later (2 to 3 days later), after the Intel upgrade tool told me so.

 

But the problem doesn't exist in 17.5 drivers since I managed to successfully repair the NVMs with their new FW on your 17.5 drivers. This was not possible with the 17.7 drivers as first noted, so the problem tracked the Intel driver.

 

Thanks for all the help in tracking down this problem. I appreciate the hard work!

 

Regards,

David

0 Kudos
Leonardo_C_Intel
Moderator
4,409 Views

Hello dmuso1

 

Thank you for your prompt response, Allow me to keep looking into this information, I will be posting back as soon as I have news.

 

Regards,

Leonardo C.

 

Intel Customer Support Technician

A Contingent Worker at Intel

0 Kudos
Leonardo_C_Intel
Moderator
4,409 Views

Hello dmuso1

 

Thank you for waiting and for sharing this behavior, as motioned previously we are working on the next release for Intel® RST, we recommend monitoring the Intel® download center and testing the new Intel® RST version when released, in the meantime, we are going to proceed to closed the thread, but if you have further questions please contact us back.

 

Hope this helps

 

Regards,

Leonardo C.

 

Intel Customer Support Technician

A Contingent Worker at Intel

0 Kudos
dmuso1
Beginner
4,409 Views

Leo ... thank you for your patience and understanding regarding this BSOD issue with version 17.7 drivers. I hope to never get another bad-disk-sector related BSOD on my computer ever but hope that should that unfortunate event happen, versions 17.8 and beyond will be able to allow me fix the problem using my disk repair tools without crashing the computer. Look forwards to versions 17.8 and beyond.

 

What would really help me is to find out if the development team verifies the problem in their 17.7 code. There has to be some unexpected exception problem in the current 17.7 code that occurs when accessing or reading/writing to bad sector sectors. Would you be able to let me know if that's the case?

 

Thank you!

-David

0 Kudos
Leonardo_C_Intel
Moderator
4,409 Views

Hello dmuso1

 

No, unfortunately, customer support does not have access to the engineering details and the information is not public.

 

Regards,

Leonardo C.

 

Intel Customer Support Technician

A Contingent Worker at Intel

0 Kudos
Reply