Slow Pre-Session Init Boot Phase (WPR Log Attached)

zinou said:

10 Gb
do you have another smaller tracer ? it will be easier for me to work with.

Sure Zinou! What components would you like me to trace?



- - - Updated - - -

I took a trace based off JustPaste.it - Share Text & Images the Easy Way

Here is the GDrive link to the new trace: DESKTOP-70EKEAN.11-18-2021.14-43-35.Boot_1.7z - Google Drive

I don't know why, but it's still pretty large at 6GB uncompressed...

Compressed file sizes are:

1. Trace 1: 9.8GB -> 469MB
2. Trace 2: 6.1GB -> 338MB

Of interest is the storport.sys calling the RAID functions, which it really shouldn't be doing. There's also the USBXHCI.SYS which seems to operate at the same time.



- - - Updated - - -

Ah, I realise that USBXHCI.SYS was not the other interesting one, it was CLASSPNP.SYS. Woops.

- - - Updated - - -

A small update:
I manage to figure out that, indeed, I had a RAID controller that did not have any drivers. It was showing up as a generic SATA device, but when I checked the device VID/DEV I realised that it was a 'AMD-RAID Bottom' device. So I downloaded an extracted driver and pointed the SATA device to it, which it happily accepted. I then loaded drivers for AMD-RAID into the driver store, even though there was no device available yet. Upon restarting, a new device appeared called 'AMD-RAID Controller [storport]' which had hooked onto the drivers I previously loaded.
After doing this I was able to successfully boot into RAID mode and install the AMD RAID drivers (it would fail before!). Problem with that is, now I cannot boot into AHCI mode, it just BSODs, but that's not really a concern for now. The concern is that the boot delay is still present even with RAID enabled. The driver behaviour looks quite different now as well, but it still seems to be getting stuck somewhere.

Here is an image of the new trace:


Here is the link to the new trace:
DESKTOP-70EKEAN.11-19-2021.13-43-10.Boot_1.7z - Google Drive

When you switch BIOS to AHCI mode, you must first boot into safe mode - did you try that?

steve108 said:

When you switch BIOS to AHCI mode, you must first boot into safe mode - did you try that?

I did not! Booting to safe mode allowed me to delete the RAID drivers in AHCI mode and I successfully returned to ACHI without safe mode. Thanks for reminding me!

The boot issue is still there though

zinou said:

I think your problem is caused by a bad snapshot of the C: driver



I think you have the same issue reported in this Thread:

Insight wanted to analyse slow boot with WPA / WPR

To fix the issue run this command:
vssadmin delete shadows /for=C: /all

When I tried to check the saved shadows with vssadmin list shadows it took approximately 20s to return the query results, so my hopes were high. However, after deleting, the boot delay was still there in all its glory.

I have taken another trace with the shadow volumes deleted, I have uploaded it here: DESKTOP-70EKEAN.11-20-2021.03-17-33.Boot_1.7z - Google Drive

Thanks for taking the time to look at my WPR files Zinou

I'm still confused as to why the function that checks the existence of RAID devices is returning a non-zero value...

zinou said:

Are you disks actually in RAID config?

If not have you tried to remove the 1 TB SSD disk for test?

- - - Updated - - -

How many disks do you have ?

I see in the boot trace 4:

1. Samsung SSD 860 EVO 4TB
2. Samsung SSD 970 PRO 1TB
3. Samsung Flash Drive FIT USB Device
4. Unknown disk !

the 3rd disk is a usb drive.
The 4th one is unidentified. There is a warning on that disk: "WARNING: Disk configuration information is invalid."

On the Samsun 970 1TB disk there is no partition! is it the disk that was used in the RAID cluster?

The CPU graph shows that the CPU is busy for 19s trying to mount a volume !



Whatever it is, the issue is related to a disk driver or a disk misconfiguration !

I suggest you unplug the USB drive, and remove all the other disks and boot your computer only with the Evo 4TB disk.

The 4th disk is just the weird Microsoft Xvd disk that Windows Store installs when you download games from the Xbox Game Pass:


In terms of disconnecting drives, I think the issue was there prior to the installation of the 970 M.2 drive, but I will give disconnecting everything a go and do another trace.

There had been no RAID configuration on the system when the issue started. I did try to install RAID and boot using RAID, based on the info in this thread, but that did not resolve the issue.

When I was going through the logs I, too, noticed that the NtfsScanEntireBitmap seemed to get stuck. The issue started when I changed from an Intel CPU/Mobo to an AMD CPU/Mobo. It could be the case that something got corrupted in the transition, but regular repairs aren't cutting it. That particular section in NtfsMountVolume spends a LOT of time in storport.sys!Raid* functions which, as I mentioned before, should not be getting called if no RAID devices exist on the system based on the decompiled storport.sys code.

I will try to disconnect all of the drives and boot again. It's possible that the BIOS is reporting the 970 as a RAID candidate for some reason. Based on your analysis, would it be worth doing another WPR log with all of the extra drives disconnected if the issue persists?

If the issue is related to the change of the mobo then an in-place upgrade of windows can fix it.

I saw a case where a user was experiencing a long boot delay after replacing his disk.
The delay has gone after an in-place upgrade.

Windows will reinstall the good drivers based on the new hardware configuration.

@zinou,
OP lists in-place as 4th item tried

zinou said:

If the issue is related to the change of the mobo then an in-place upgrade of windows can fix it.

I saw a case where a user was experiencing a long boot delay after replacing his disk.
The delay has gone after an in-place upgrade.

Windows will reinstall the good drivers based on the new hardware configuration.

I actually did an inplace upgrade recently but it didn't resolve the issue. It has also persisted through a full reinstall (without format though). I'm leaning to it being a bug in the bios, and I'm tossing up whether it's worth hex editing storport.sys to bypass that function call since I'm not going to need RAID anytime soon.

- - - Updated - - -

Sieyk said:

I actually did an inplace upgrade recently but it didn't resolve the issue. It has also persisted through a full reinstall (without format though). I'm leaning to it being a bug in the bios, and I'm tossing up whether it's worth hex editing storport.sys to bypass that function call since I'm not going to need RAID anytime soon.

Hex editing failed. Even with early launch antimalware disabled and no driver verification, it fails instantly. I guess I shouldn't pretend to know how to patch system code without introducing errors