Anyone Else with Troubles following the October Windows Updates for Win 7?

[BonzaiDuck]

Concerning a very solid running-gaming-working-sleeping-hibernating-waking system with 2x GTX 970 in SLI. And it was overclocked -- the 2700K in my sig well-within the best wisdom of the SB enthusiast community for voltage and speed. The RAM -- 20 GB G.SKILL DDR3-1866 9-10-9-28 -- was run at spec, had all passed 500% HCI Memtest64 when I augmented the 16GB 2x8 kit about three months ago. It's now running at stock speeds, and I'm running HCI Memtest again -- just to rule out hardware. I may remove a graphics card next, if I can disable SLI from a Safe Mode session. It seemed stable in Safe Mode.

Suddenly, and at first with the Win 7 system only, it acted as if the boot disk was corrupted. But the boot disk, for the testing, seems to be stellar.

I restored the entire system from the October 4 nightly WHS-2011 backup. The system was "solid" again. I ran Steam games, watched encrypted TV from my SillyDust tuners, poked around on the web, did a software inventory because I was transferring programs from the 2600K when this happened. Ran IntelBurnTest through 20 iterations, OCCT Linpack for a half hour. Rock solid. I might have run a graphics stressor, but the gaming would have seemed enough.

Then, the notice under Windows Updates reminded me that I had to do them over because this was the October 4 backup of the OS.

Suddenly, after they completed, my misery returned. I suspect an "optional" update as the cause.

I figure if anyone else is suddenly having troubles, it will help me to know if there are independent observations with the same suspect indications. That is, there is the "anecdotal," "the statistical," and the "digital." With "digital," a smaller set of sample events should prove something.

I can only say I've been through this cycle twice, all resulting from the October Windows Updates (and a choice to include some "Optional Updates.) I'll complete the second pass when I reinstall Windows from the server.

. . . And I think the cause involves some interaction between my current hardware drivers and those updates. Further, the 2600K twin of the troubled system never got those optional updates. It also doesn't have SLI, and only uses a single 2x RAM kit.

 

[corkyg]

My 7 got 'em all last night, and I added an optional. No problems whatsoever.

 

[mikeymikec]

No problems, updates installed yesterday and this morning.

 

[BonzaiDuck]

No problems, updates installed yesterday and this morning.

Also, I don't think I mentioned this. You and other responders would know by now. This month included an "update rollup." The "don't turn off your computers . . . . configuring updates" at both shutdown and the subsequent startup seem to take so long, anybody could think they've "hung." Between that message screen with the rolling-circle of a mouse-cursor, and the user log-in screen, some folks may see a black computers screen -- completely blank -- but for a mouse-cursor that responds, but can't move you forward. With such a screen (and the delay has occurred for more than one of my computers), a user could think things have "gone south," and choose to hit "reset" or do a cold shutdown. All of these things have risk-potential for fouling up windows.

I really appreciate responses to the thread. More information if anyone has had my experience: 2x GTX 970 in SLI, and I think the latest NVidia driver. Also, the Z68 board is running a built-in Intel LAN. Symptoms might suggest there's something wrong with video stability, and the internet icon in SysTray shows the rolling circle, unable to complete the connection.

Called my friend the retired electronics tech in Virginia. He agrees: Likely a software and driver conflict arising from the updates. It could still be hardware -- that's why I'm testing. So far, the RAM testing has passed the "quick tests" and 22% through HCIMemtest64.

But simply restoring a WHS backup from 10 days ago makes it all "go away" until I reinstall those Windows updates. I may attempt to install them one at a time, until I can locate which might be causing the trouble.

Before that, I want to restore again from the October 4 backup, restore the default clock settings on the 970 cards, and disable SLI. Possibly also -- update the LAN driver.

Bill Gates and his corporate entourage have all been replaced by business graduates who don't believe in cross-subsidizing service and product from effectively monopoly profits.

 

[MustISO]

I've done the updates on 5 Win7 x64 systems and no problem at all.

 

[escrow4]

Old overclocked CPU that likely needs more juice to keep going now, SLI and mismatched RAM is not solid. Its a slapdash amount of hardware that might or might not run properly and something in the latest batch of updates for creaky old corpse 7 is making it take a dump. And despite what everyone else says an overclocked system is never truly stable, its always on the cusp of falling apart vs running faster than stock.

 

[BonzaiDuck]

Parts are only 2 years old.

I'll surely update this thread if I find it was a video or LAN driver conflict. With something such as Escrow4 suggests, it would not be working fine, take a dump after the updates, and then work fine for a solid day's time of work-out use, then take a dump after the second round of updates.

Processor was never overvolted -- load and idle voltage range was 1.344 to 1.38V, either at 4.6 or 4.7 or simply 3.9Ghz at lower voltage as it is now.

Since RAM is on the verge of passing a 100% HCIMemtest64, it's either a dying graphics card or a driver conflict -- that's what I think. And the graphics cards weren't "abused." Never got hotter than 65C except during stress, when it would max out at 70C.

the behaviors are simply too robust and discrete to think it was anything other than drivers, the graphics cards or the LAN. It wouldn't simply be "hard disk corruption," if I could restore the system to the same hardware, check the hardware, run the system for 24 hours through its paces, then reinstall updates and have it go south.

If it's actually a HARDWARE problem with the graphics, likely one card is still good, and I can still "deal with it."

Next step is to restore (again), enter Windows, shut off Windows Update temporarily, uninstall the graphics drivers after resetting the cards through AfterBurner, then uninstall Afterburner.

Even NVidia and Afterburner would show a glitch before, which would get fixed with finality, then updates would bork it and I had to uninstall then reinstall both the driver and AB. the glitch would occur for attempting to install the driver with SLI enabled: and it would affect Afterburner and the power-states of the card -- which would not settle back to their 135Mhz level after a workout from a fresh boot. This was always fixed in the same manner: reset, uninstall AB and driver, reinstall carefully, and -- fine.

But -- YES -- I'll update with my findings here.

 

[Ketchup]

When 7 got near the end of its life on my desktop, I found myself skipping most of the optional updates. They were all to either update hardware to an older driver than I had already, or improve the Windows 10 upgrade experience, which I had already done on anothet drive.

But to answer your question, I ran the update on a Lenovo laptop with Nvidia graphics and a VM running 2 cores of my 4790, and both behaved fine all day.

 

[BonzaiDuck]

When 7 got near the end of its life on my desktop, I found myself skipping most of the optional updates. They were all to either update hardware to an older driver than I had already, or improve the Windows 10 upgrade experience, which I had already done on anothet drive.

But to answer your question, I ran the update on a Lenovo laptop with Nvidia graphics and a VM running 2 cores of my 4790, and both behaved fine all day.

thanks. I think if I follow my plan, I'll find one of two things:

-- a driver conflict with the updates
-- a failing graphics card, or marginal OC setting.

There should have been more than adequate voltage for the modest OC, which was 1480Mhz core and 7500Mhz memory. it may have required a 20mV increase with the power limit at 110%, temperature limit at 85C (but it never exceeded ~76 during stressing.)

Yet cannot explain retrieval of total stability with a workout over some 24 hours, the IntelBurnTest results and so on, except that it suddenly took a dump again after the Win Update installation.

I never use MS's WHQL drivers for hardware, and always go to the mVidia site for my graphics drivers.

This system is probably as unique as any other NVidia-graphics configuration. But as of now, I cannot explain it with RAM hardware failure or CPU failure. If anything failed, it would be one of the graphics cards. But it would not restore from a 9-day-old backup to total stability if it were a hardware problem. That doesn't make sense.

With or without OC'ing, there is always this possibility of confusing a hardware failure with a software and driver problem. I'll just have to see it through.

All my files are safe! I still have server backups of the system! Things could be a lot worse.

 

[BonzaiDuck]

OK -- doesn't seem to be the hardware, although I set the clock back for the SLI GTX cards.

There was some sort of problem with the dual-boot configuration. Since I can reinstall it and I wasn't using it much except to experiment with replacements for WMC, I deleted it.

So far, things look tip-top until I install those updates.

I'm inclined to wonder whether I should uninstall both afterburner and the graphics driver, THEN install the updates, then re-install the driver and afterburner, and all of this done with SLI disabled.

Gotta think . . .
[LATER . . ] Mmm. This could be more a "Disk & Storage" issue. Without yet going forward to download the latest Win Updates, this is now a stable single-boot Win 7 system. Lost nothing for deleting the Win 10 part via "MSCONFIG->Boot[tab] and purging the drive volumes adding to "unallocated."

For some weird reason, the [dual] boot-volume for C: had been converted to "Extended partition" and "Logical." On the twin system -- totally stable 2600K -- the equivalent partition or win 7 C: drive was simply Primary. The system stub of 300+MB created by Win 10 was at the end of the disk for the troubled system, and in sequence following the [Win 7] system-reserved and preceding C:.

And -- there's more -- when I set WHS-2011 to back up the system, I can only guess that dialog had cut off the scroll list and I hadn't looked at the entire thing: I thought that the drive only had "reserved" + C:Win7 + Win10. So I'd never backed up the other 300+MB reserved volume in the nightly backup. What was never restored, continued to stay there. These last couple days are the first and only occurrence of a "restore" from the server.

Should I post this as a thread on "Storage & Memory" or simply leave it here?

Stuck in my old Acronis DD11 update-2, and changed the appropriate items, also resizing the volumes for time-being without Win 10. If this system has to do WMC duty, I may just leave it as a single-boot until I'm sure how it got fouled up after some six months rock-stable as dual-boot.

 

[corkyg]

I might add, that as a general policy, once an update has started, for me it is hands off except for the reboot. Then the usual "Do Not Turn Off" stuff happens, and there are times when the video is black. I expect that and do nothing.

 

[BonzaiDuck]

I might add, that as a general policy, once an update has started, for me it is hands off except for the reboot. Then the usual "Do Not Turn Off" stuff happens, and there are times when the video is black. I expect that and do nothing.

The plot thickens.

I had restored the troubled computer back to October 8. rock solid. Set the remainder of OC (SLI GTX 970 x2) back to default. Left SLI on. didn't touch the driver and afterburner.

Then began installing the October updates one or two at a time.

finally, it came to Silverlight 5 -- from a "50709" Sept8/Sept12 version, to maybe "909" -- the update that was in the October set.

Not only did that unravel the troubled machine that had been put back right, I also went to the 2600K machine to do a routine reboot (usually can go days or weeks between). When it came back, same sort of symptoms I'd had with the 2700K, and I had allowed all the updates -- perhaps even that machine simply installed without user intervention in the background.

The first troubled machine -- the 2700K -- so far has "come back" by rolling back the system to a previous restore point. that was quite handy, because I had been installing updates one at a time.

This is beginning to seem like MS's "contribution" to the cyber-warfare and security threat. For all I know, it might have something to do with whether "Secure Boot" is enabled. As it stands, I can't get past the front door of the Microsoft Download/Update Catalog on that system. Maybe it has something to do with the Silverlight rollback. No problem on the new Z170 system: which I used to download the previous Silverlight in case the rollback leaves it in a flaky state, which I'm beginning to suspect.

I'm just speculating about all this. I could only make an intelligent guess that anyone else with this problem has this situation:

1) They use the Silverlight plug-in
2) They have a prior gen mobo that doesn't include the Secure Boot feature
3) They have Win10 on top of Win 7 in a dual-boot
4) My 2600K system has single NVidia 780, and the 2700K has 2x 970.

The combination of (1) and (3) may be more relevant to stable system and software without malfunction. The first "troubled system" had Win10 excised, then became "troubled" again after the Silverlight October download -- a robust and discrete difference over what had been restored to rock-stable, able to pass Affinitized LinX (stock clock) 15 passes like a speeding bullet and leaping over tall buildings etc. The second one is still in a dual-boot state, and I was able to get into "Safe Mode" without totally trashing the OS and doing a restore on a SECOND computer from the WHS server.

I cannot guarantee that the three conditions or points are applicable, but this involved two dual-boot Win7/Win10 systems, and the trouble arose from both with the latest Silverlight download.

I'm not even sure that everybody has Silverlight installed or uses it, but it's just a plug-in for IE.

Who's pushing people toward total dependence on Win 10 without Media Center? Microsoft is.

Who is pushing Silverlight in Win Updates? Microsoft is.

Who's part of the consortium with the newer "Secure Boot" UEFI feature? Microsoft is.

Is there going to be dissent from the peanut gallery when I say again that MS has wandered downhill after Gates and the others left or went on to other things?

Shame on MS! I suppose, anyway -- unless there's some peculiar reason that I've had this problem and nobody else does, or that I've somehow mis-diagnosed multiple-causation.

I'm pretty sure now that this has nothing to do with OC clocks, and nothing to do with SLI and the graphics driver, unless the new Silverlight somehow conflicts with the most recent NVidia driver or at least the one behind it (they seem to issue those things monthly!)

I also heard that NVidia and MS are having some feud. NVidia had previously had a spat with Intel some few years back. The latest skinny is just gossip. I don't know how accurate it is, or how it bears on this . . . . . trouble.

DO I FREAKIN' EVEN NEED SILVERLIGHT?! What are the implications of uninstalling it from INternet explorer? "Edge" in Windows 10 doesn't even use it.

Maybe if I whine enough and threaten to go to the county social service worker who will ask me to "touch the dolly," Microsoft will atone for its crimes . . . .

ADDENDUM: (and sorry for the length, but this mushrooming glitch has made me crazy).

Add one more facet of possible multiple causation: STEAM.

 

[Ketchup]

I stopped needing Silverlight when Netflix started supporting HTML5.

 

[BonzaiDuck]

On the two machines that I have "fixed" back to stability, I supposedly have the older version of Silverlight installed, but assuming a possibility posed by the warnings about rolling back to a prior checkpoint or restore point, I may have to install that older version. I was able to get it through the SKYLAKE-in-Progress system, but now, when I go to MS Updates Catalog to download, it tells me to "enable User Persistence" in IE options. But it IS enabled!

Are you able to get into that Catalog webpage WITHOUT Silverlight? If I don't NEED Silverlight, I don't want it, but now I won't tolerate it if I don't need it.

[UPDATE:] reinstalled the earlier Silverlight, then discovered what needs to be done with "Internet options" in IE to get into the Updates Catalog.

OK. Puts that to rest.

I DIDN'T DO ANYTHING WRONG! Been working with Win 7 since 2011. Dual-boot was working fine these last 5 months. Not so much as an artifact, BSOD, "Program stopped working" -- none of it.

I SAY -- the gov'mint needs to "regulate' MS while cooperating on the issue of security.

this . . . is all quite . . . .in SANE.