Windows corrupts after latest patch corrupt primocache Topic is solved

[brighttail]

At this point i think I'm done with attempting an L2 cache for the time being. For me it isn't worth the hassle.

I did notice in one of my many resinstallations that there was an issue with the GPU drivers. I don't remember the exact error but i was BSODing at startup and managed to get into safe mode to check the logs. It was a driver issue. I have an NVIDIA uninsntaller that gets rid of EVERYTHING from within safe mode and did that, reinstalled the drivers and was able to get back into windows.

That issue tho was completely separate from the scanning/repair issue as that only happens now when I've got an l2 cache going.

So I give up for now, not going to deal with it anymore as it frankly isn't worth any more time.

[Support]

I'm really sorry for the late reply as we're busy in testing and fixing for v3.0 and Win10 1703build!
Regarding the message "Scanning and repairing the volume..." during the boot-up in Win 10 1703build, we already have a fix for this problem. The cause is that PrimoCache L2Storage volume is a dedicated/private volume and Windows cannot correctly recognize it. So when you have a L2Storage partition, you will see such kind of message.

[brighttail]

I'm really sorry for the late reply as we're busy in testing and fixing for v3.0 and Win10 1703build!
Regarding the message "Scanning and repairing the volume..." during the boot-up in Win 10 1703build, we already have a fix for this problem. The cause is that PrimoCache L2Storage volume is a dedicated/private volume and Windows cannot correctly recognize it. So when you have a L2Storage partition, you will see such kind of message.

That is good news to hear. So you said you have a fix for the solution. I'm assuming that will be done in v3.0?

The problem with that scanning/repair message for me is after a few times, windows believes something is broken and will go into 'repair' mode until that partition is deleted.

I really do think Microsoft made this issue on accident (or maybe on purpose) getting their build ready for the Optane drive which would be a separate, physical drive.

[MaddBomber83]

Just to throw in; a few questions posted asking if everyone was a magician Samsung type. I am not. OTZ RD 400 1T / WD Blue 5T.
I tried mine with both MS and Toshiba Drivers.

I also ran through this with Toshiba support before even thinking of coming through here... So a secure erase and fresh install resolved the volume message (probably because it got rid of the Cache Volume). Just a note for the future... I took a video of my startup and extracted the frame that contained the full message.

After reading this, my time with Toshiba was almost wasted; we found an overheat issue that is now resolved.

Then without installing anything else except the normal updates, I installed what I believe was the 2.7 version. When rebooting I was given the startup corruption deal; something windows could not recover from. I could however go back to a restore point. But only once. I should note that no tasks or partitions were made at this time; just the program being installed was enough to cause corruption. I went so far as to install windows on my HDD and disabled the x4 drive via bios. Same deal; boot corruption.

;;;;;;;;;;;;;;

It looks like another attempt is in order. From what I've read there was no real change to how people did this on a fresh install.

[BonzaiDuck]

I'm really sorry for the late reply as we're busy in testing and fixing for v3.0 and Win10 1703build!
Regarding the message "Scanning and repairing the volume..." during the boot-up in Win 10 1703build, we already have a fix for this problem. The cause is that PrimoCache L2Storage volume is a dedicated/private volume and Windows cannot correctly recognize it. So when you have a L2Storage partition, you will see such kind of message.

Well, I didn't download a patch for your ongoing and very important project there! So how did I get rid of it?

There are many posts following yours in recent days, and I haven't even had time to read.

Here's something else that is interesting, but I'll just provide the link to my long and lonely dissertation about it:

https://forums.anandtech.com/threads/so ... t-38872334

"Everything (for me) works OK -- as long as I don't do certain things." And I know you'll say "Well! Don't over-clock your f[bleep]-in' computer!"

So not to blame Romex or PrimoCache -- Windows 10 is common to the complete array of symptoms discussed so far.

This one? YOu can save yourself some trouble by telling me "we won't investigate problems on your system if it's overclocked." What would I do, in your place and your place of business? Your choice. Somehow, I think this relates to memory management, Windows, PrimoCache and some other things. I don't rightly know. But I think I have it licked. Or rather I know how to avoid it. Have to wait and see until I see it passing the current test for another 90 minutes . . .

[BonzaiDuck]

Just to throw in; a few questions posted asking if everyone was a magician Samsung type. I am not. OTZ RD 400 1T / WD Blue 5T.
I tried mine with both MS and Toshiba Drivers.

I also ran through this with Toshiba support before even thinking of coming through here... So a secure erase and fresh install resolved the volume message (probably because it got rid of the Cache Volume). Just a note for the future... I took a video of my startup and extracted the frame that contained the full message.

After reading this, my time with Toshiba was almost wasted; we found an overheat issue that is now resolved.

Then without installing anything else except the normal updates, I installed what I believe was the 2.7 version. When rebooting I was given the startup corruption deal; something windows could not recover from. I could however go back to a restore point. But only once. I should note that no tasks or partitions were made at this time; just the program being installed was enough to cause corruption. I went so far as to install windows on my HDD and disabled the x4 drive via bios. Same deal; boot corruption.

;;;;;;;;;;;;;;

It looks like another attempt is in order. From what I've read there was no real change to how people did this on a fresh install.

I think I posted my observations about how to get the 1703 build -- one approach being better than the other. Go through Windows Update. I think it will get you through the KB. . . 240 installation, from where you can first make a USB or ISO customized to your hardware. If you go to the link which begins to update the W_U_ Assistant (KB . . . 240) and but back out of the choice to install, it could bork Win Update operation showing stuck at 0% on the KB/Assistant update -- get the link to that page and don't lose it. I was able to pick up from there.

I mentioned some of my thoughts about a nexus of causes for a handful of effects. I'm going to shut my yap at this point and sit back and watch what develops. I'm probably the only person here who actually thought this misery was some type of fun for me.

Also! Nearly forgot! Some tech forums tend to be a magnet for one type of question or concern, and other forums draw a different or even robust set.

Microsoft released an advisory (somewhere) mentioned recently at Anandtech -- check the Operating Systems forum. Build 1703 should be approached with caution until your Build 1607 WU advertises a notice about it in your primary WU dialog. They say it is "ready" for some groups of users and not others. There are "bugs" to fix which may apply to some and not to others. Somehow, inferring from my previous paragraph, I'd wager that folks here are more likely to have received the notice in their WU window. The other people are enthusiast hot-dawgs who regularly download and install pre-release Beta versions, so they would've discovered it first. The might have attempted to download it from an advance pre-release link, or to use the IT admin "full" download.

The more I talk about what I think I know, the more I think I should be mum for a while.

[brighttail]

As far as I can tell, tech support here hasn't said anything about overclocking or such, they have simply acknowledged that there is/was an issue with the dedicated/private volume that Primocache sets up. Windows cannot correctly recognize it.

Unfortunately for me, at least, the new version 2.7.2 hasn't corrected this issue with the L2 Storage volume so I'm hoping the fix comes out in v3.0. Until then my solution is simply not to use it.

[BonzaiDuck]

As far as I can tell, tech support here hasn't said anything about overclocking or such, they have simply acknowledged that there is/was an issue with the dedicated/private volume that Primocache sets up. Windows cannot correctly recognize it.

Unfortunately for me, at least, the new version 2.7.2 hasn't corrected this issue with the L2 Storage volume so I'm hoping the fix comes out in v3.0. Until then my solution is simply not to use it.

That sounds like the reasonable and wise approach! On the OC issue, it was something I couldn't help doing since the Northwood socket LGA-478. It's sort of like the JFK assassination and attitudes about devoted researchers: you're "out-of-spec" and doing something that isn't guaranteed by your hardware vendor -- you are suspected to be flakey. But the "K" chips are shipped just with that possibility in mind, and those hard-core veterans (me?) are throwing everything they can at these "out-of-spec" settings to make sure they're good. People were told that Prime95 or LinX would damage their processor, but that would only happen through overvolting the CPU or overheating it. I can run those tests and fall asleep knowing that they'd only throw an error they can manage. So to some OC enthusiasts, I'm a sissy.

But I've got my 4.7 Ghz back! Rock-solid! [UPDATE: 2hr 40min OCCT:CPU and counting. Looks like I've got this licked . . . needs more tests, but it looks good. ] [Later:] Yes, with the original 4.7 VCORE of 1.408V I'd found last fall, we're back, we're bad, we're on top. For ordinary gaming and real-world applications, the VCORE never exceeds 1.392V. Z170 boards are often shipped to default the VCORE to as much as 1.40V. Infer the secret conversations between INtel and "Board partners." Three stress programs pass: Prime95, IntelBurnTest, OCCT:CPU. Trouble with LinX last year's version -- I forgot to configure threads and cores when I installed it on the Win 10 OS. That FAIL is turning into a PASS. All this complicates the other problems, except that I solved the other problems at lower clock settings, and they don't reappear now. I only wish others here had better luck.

But Primo Support is on the job!

[Support]

We now have made a patch for the L2Storage volume problem in Win10 1703build.
You may download the latest version 2.7.3 from the following page.
https://www.romexsoftware.com/en-us/pri ... nload.html

And here is the change log.
https://www.romexsoftware.com/en-us/pri ... gelog.html

For existing L2Storage volumes, I'm sorry that you have to re-create them with the latest version to solve the error message.
Thank you!

[BonzaiDuck]

We now have made a patch for the L2Storage volume problem in Win10 1703build.
You may download the latest version 2.7.3 from the following page.
https://www.romexsoftware.com/en-us/pri ... nload.html

And here is the change log.
https://www.romexsoftware.com/en-us/pri ... gelog.html

For existing L2Storage volumes, I'm sorry that you have to re-create them with the latest version to solve the error message.
Thank you!

That symptom I somehow expunged before your patch, but I'm going to delete and recreate my L2 volume after installing v.2.7.3.

I may have noticed some other behaviors, but they are interactions possibly between L1 caches and other software that can grab as much memory as they possibly can. Such software is:

Prime95 v. 28.x
LinX v.0.6.9
OCCT [:CPU] -- latest version

Nothing seems to go wrong with the caches, but this software throws unreliable results. Once the L1 caches are deleted (possibly even "paused"!), the errors go away. The software is used to validate overclock settings, and you will no doubt be skeptical about my observations, or maybe assume that I've fouled something up with BIOS "out-of-spec" settings -- but no. It seems to be a conflict involving memory management.

You could look into it, but if you choose not to address it, that is also not a real problem. One could simply warn "Don't run these types of programs with active caches, and define the caches after you're finished using those programs." Even so, I observe that the same programs work properly with active L1 caches testing the same BIOS settings under Windows 7.

I am really looking forward to version 3.0, which you had suggested might allow benchmark programs to show true performance under L2.

Also, I made an observation in another thread about concurrent L1/L2 cache configurations pertaining to disk errors I'd occasionally received, needing CHKDSK repair. Eliminating the L1 component for an L2-cached HDD seemed to fix it. The HDDs I'm using are 2.5" Seagate "laptop" drives: A Barracuda 2TB and a 2TB "Seagate-Samsung."

Again -- thanks for your attention to myself and our colleagues here concerning the Windows Creator Update Crisis we all shared.