Rare thirty-seconds OS freezes related to NVMe SSD

[vishalrao]

I’m running Win11 Pro on a new 9970X CPU + ASUS TRX50 (older non-A variant) board with a Samsung 9100 Pro SSD and a SuperFlower Leadex Platinum ATX 3.1 1000w PSU.

Very rarely the system will freeze for almost exactly thirty seconds then go back to normal.

Upon some web-searching and looking at event viewer, there’s event IDs (forgot the numbers) related to NVMe (“stornvme.sys” IIRC) saying something like “write blocked/failed” about the same time the issue occurs.

Searching for the issue, some similar reports online suggesting defect with either the board, SSD (firmware issue or just defect) or in some cases PSU. Suggested workaround to enable a Windows power saving related setting via registry edit something to do with “AHCI sleep” which I tried and it didn’t help.

Does anyone have any pointers or suggestions about this issue? Or should I start replacing the parts one by one?

Side note that my board is really old inventory (manufactured over a year ago!) which makes me want to replace it first before the other parts.

Thoughts?

Oh what about any BIOS settings to try? There’s some power related option IIRC like “idle current typical” or “idle blah” gosh I should really look it up and write it properly here…

 

[bobbyprajan]

Have you checked the logs at Event Viewer > System ? Entries with Error type can help you to troubleshoot this further. Also try turn off all power management features in BIOS/UEFI to see if it is related to that

 

[rwnano]

Couple of things
You are in Pune , MBhanghui sir is also in Pune.
I suggest you try using Ubuntu Linux on your PC. Much stabler experience than Windows

 

[vishalrao]

Have you checked the logs at Event Viewer > System ? Entries with Error type can help you to troubleshoot this further. Also try turn off all power management features in BIOS/UEFI to see if it is related to that

Thanks, yes, I have seen relevant event viewer errors related to this, plus tried changing a windows side hidden power setting, but issue recurred again.

And yes, my next step is to change a power related setting in BIOS to "idle current typical" instead of whatever is the default idle setting, it's somewhere in there.

First, I have been trying to trigger the issue on-demand to no avail, which is infuriating because if I change something then have to wait for an unknown duration to see if the issue is resolved by that change, but if it does not happen, I don't know if it is resolved or not really

Thanks again. I'll keep digging!

Couple of things
You are in Pune , MBhanghui sir is also in Pune.
I suggest you try using Ubuntu Linux on your PC. Much stabler experience than Windows

Yes, I have a few linux installs alongside Win11 too, just that I'm mostly running Win11 for my work, only sometimes use linux, on top of the fact that this issue only occurs rarely (and I bet won't occur on linux hehe) I'll need to be patient in trying to resolve this issue. Thanks!

PS: Yes I know Manvendra sir is in Pune and not too far from where I am, but not sure how that is relevant to this topic hehe. I have given him prior notice/warning that I plan to drop by his place some day to check out his audio and linux adventures first hand!

 

[bobbyprajan]

Thanks, yes, I have seen relevant event viewer errors related to this, plus tried changing a windows side hidden power setting, but issue recurred again.

You are not overclocking, are you ? If you are using XMP profiles for memory, you should try using the default profiles to see if that helps. I have had weird issues like this when I tried those. Also try disabling the power saving C states in Power Management. You can try some of the stress testing suites to see if you can trigger the fault.

 

[rwnano]

Thanks, yes, I have seen relevant event viewer errors related to this, plus tried changing a windows side hidden power setting, but issue recurred again.

And yes, my next step is to change a power related setting in BIOS to "idle current typical" instead of whatever is the default idle setting, it's somewhere in there.

First, I have been trying to trigger the issue on-demand to no avail, which is infuriating because if I change something then have to wait for an unknown duration to see if the issue is resolved by that change, but if it does not happen, I don't know if it is resolved or not really

Thanks again. I'll keep digging!



Yes, I have a few linux installs alongside Win11 too, just that I'm mostly running Win11 for my work, only sometimes use linux, on top of the fact that this issue only occurs rarely (and I bet won't occur on linux hehe) I'll need to be patient in trying to resolve this issue. Thanks!

PS: Yes I know Manvendra sir is in Pune and not too far from where I am, but not sure how that is relevant to this topic hehe. I have given him prior notice/warning that I plan to drop by his place some day to check out his audio and linux adventures first hand!

Have you tried Wine on Linux for windows programs?
Have you tried windows virtual machine tools on Ubuntu Linux?

 

[vishalrao]

You are not overclocking, are you ? If you are using XMP profiles for memory, you should try using the default profiles to see if that helps. I have had weird issues like this when I tried those. Also try disabling the power saving C states in Power Management. You can try some of the stress testing suites to see if you can trigger the fault.

No OC here. I do set EXPO for my RAM though. I've tried "stress testing" by running multiple workloads of my work (my software project builds multiple simultaneous runs) to no effect - tried Prime95 blend for over 30 minutes to no effect either - yes I have not tried disabling C-states in power management - will look into that along with that BIOS idle setting I mentioned - thanks!

Have you tried Wine on Linux for windows programs?
Have you tried windows virtual machine tools on Ubuntu Linux?

Yup tried those, it's really cool stuff indeed, but for now I'm daily-driving Windows, like I said mainly due to work needs. I've always wanted to move full time aka daily-drive Linux but finding it difficult (mainly time crunch) even though the latest WINE and KVM/QEMU releases should work well I presume Right now I run various Linux distros and boot into them regularly albeit for only brief durations to check out the latest updates and in some cases, when I do find time, actually send some code bugfixes etc, like for my two favourites KDE and elementary OS, see https://elementary.io btw if you can.