V767 BSOD (APC Index Mismatch) trying to Remove a Drive

[jmone]

V767 looks pretty good (I like the addition of the ReFS drives) but I'm having an issue when trying to remove a Drive.  I've tried a couple of time to remove a drive but both times got a BSOD with a APC Index Mismatch.  I'm now moving the files manually with the DriveServer service stopped between drives and the only thing I can see is I have some directories with Full Stops in them that Windows does not like, but I'm not sure if this is the issue.  Windows did create Mini Dumps (Attached).

Thanks

Nathan

MiniDumps.zip

[Christopher (Drashna)]

Yup, definitely a driver related issue. 

 

Though, just in case, run a memory test.   I'm pretty sure it will pass clean, but on the off chance....

 

 

Though, I'm having issues reproducing the issue.  

What OS are you using (and version number, if Windows 10), are you using any antivirus software, what options are you using, is any of the data duplicated, etc? 

[jmone]

OS: Win 10 Pro - 64Bit  - V1703 - Build 15063.448

Anti Virus: Windows Defender (std settings)

DrivePool: No Data Duplicated, Ordered File Placement Balancer, File Placement Rules 

 

Will run a memtest and report back.  Did the mini dump show anything?

 

Thanks

Nathan

[jmone]

Passed the Windows Mem Test

[Christopher (Drashna)]

The memory dump didn't show anything useful.  To quote Alex: 

 

Unfortunately, the driver and code that caused this is long gone, so there's no way to know where the problem is. 

 

 

If this is continuing to occur, then enable the driver verifier and repeat. 

 

To enable the driver verifier, run "verifier" on the system.  Create a new profile, select a specific driver, and select "covefs.sys".   Click through and reboot the system. 

 

Once that's done, try removing the drive.   

 

If it crashes, it should give a more detailed dump, and may contain the information that we'd need to troubleshoot this properly. 

 

Once you're done, run "verifier" again, and delete the profile. 

[jmone]

No probs - I'm doing a bunch of shuffling (that will take a week or so) so I can reformat as ReFS V3.  I'll then get to remove a couple of drives so I will see if there is any error.  Out of interest what did ALex mean that the driver and code is "long gone".  Do I have some old component as it was on V767.

[Christopher (Drashna)]

That the info in the memory dump is "long gone". As in, it's been unloaded from memory already, so no way to see what caused it. 

 

The driver verifier is .... more aggressive about logging, and error detections, so it should help identify issues that wouldn't normally show up, or capture them "sooner".