After powering on PC daily, its W10 seems to always seem to show :( about it running into a problem...

For a few days, I have been using the client's Intel NUC PC with its W10 b1803
weird issues. Originally, I was asked to look at why its W10 takes so long to
start/boot up. It seems to happen on each new day after power it on when it was off
for many hours like overnight. I can't seem to reproduce it after it does its scan
and repairs. I even tried changing BIOS' date to a day later to see if I could
reproduce it. Nope.

I wonder why it is always happen daily. Failing hardwares somewhere? In W10, I ran
W10's disk checks on both EMMC and SSD, and it found no issues. I know the issues
happened on both drives (original small EMMC and cloned bigger SSD). So, I decided
to force a chkdsk /r /f on both C: and D: drives outside of W10. C: was quick, but
D: took a while especially at 45% with intense light activities that long.

After returning to normal W10 session, I checked its Event Logs' Applications
section with its filter to show only 1001 event ID
(https://winaero.com/blog/how-to-find...n-windows-10/). I exported
and uploaded to http://zimage.com/~ant/temp/W10appEventLogs1001.txt. There were
lots of chkdsk repairs on both drives, crashes, etc.

What do you guys think is happening?
--
Quote of the Week: "As a thinker and planner, the ant is the equal of
any savage race of men; as a self-educated specialist in several arts
she is the superior of any savage race of men; and in one or two high
mental qualities she is above the reach of any man..." --Mark Twain
Note: A fixed width font (Courier, Monospace, etc.) is required to see this signature correctly.
/\___/\ Ant(Dude) @ http://aqfl.net & http://antfarm.home.dhs.org /
/ /\ /\ \ http://antfarm.ma.cx. Please nuke ANT if replying by e-mail.
| |o o| |
\ _ /
( )

I think you need to urinate in the president's mouth.

On 2/21/2019 4:24 AM, Ant wrote:
For a few days, I have been using the client's Intel NUC PC with its W10 b1803
weird issues. Originally, I was asked to look at why its W10 takes so long to
start/boot up. It seems to happen on each new day after power it on when it was off
for many hours like overnight. I can't seem to reproduce it after it does its scan
and repairs. I even tried changing BIOS' date to a day later to see if I could
reproduce it. Nope.

I wonder why it is always happen daily. Failing hardwares somewhere? In W10, I ran
W10's disk checks on both EMMC and SSD, and it found no issues. I know the issues
happened on both drives (original small EMMC and cloned bigger SSD). So, I decided
to force a chkdsk /r /f on both C: and D: drives outside of W10. C: was quick, but
D: took a while especially at 45% with intense light activities that long.

After returning to normal W10 session, I checked its Event Logs' Applications
section with its filter to show only 1001 event ID
(https://winaero.com/blog/how-to-find...n-windows-10/). I exported
and uploaded to http://zimage.com/~ant/temp/W10appEventLogs1001.txt. There were
lots of chkdsk repairs on both drives, crashes, etc.

What do you guys think is happening?

Not sure but I'd consider replacing the motherboard battery since they
are relatively cheap. Not that we know the age of the existing battery.

I'm sure Paul will have something that will nail it down for you when he
gets around to reading the newsgroup.

Ant wrote:
For a few days, I have been using the client's Intel NUC PC with its W10 b1803
weird issues. Originally, I was asked to look at why its W10 takes so long to
start/boot up. It seems to happen on each new day after power it on when it was off
for many hours like overnight. I can't seem to reproduce it after it does its scan
and repairs. I even tried changing BIOS' date to a day later to see if I could
reproduce it. Nope.

I wonder why it is always happen daily. Failing hardwares somewhere? In W10, I ran
W10's disk checks on both EMMC and SSD, and it found no issues. I know the issues
happened on both drives (original small EMMC and cloned bigger SSD). So, I decided
to force a chkdsk /r /f on both C: and D: drives outside of W10. C: was quick, but
D: took a while especially at 45% with intense light activities that long.

After returning to normal W10 session, I checked its Event Logs' Applications
section with its filter to show only 1001 event ID
(https://winaero.com/blog/how-to-find...n-windows-10/). I exported
and uploaded to http://zimage.com/~ant/temp/W10appEventLogs1001.txt. There were
lots of chkdsk repairs on both drives, crashes, etc.

What do you guys think is happening?

Is the computer in question near an ionizing radiation source ?

What about power quality ? Where the NUC is plugged in, is
the power clean ? Could the problem be "injected" noise coming
from noisy machinery with inductive spikes ?

I see it is scanning a file system at rest. I could understand
scanning the C: drive, which is an active partition on each session,
and an inproper shutdown (drive isn't flushing properly), could
leave orphan this or that.

But the distribution of errors is a bit weird. It doesn't
look entirely like dirty shutdowns. It's like storage
just isn't working properly.

Since you have a new SSD and it's being checked too, I'd have
to guess that the +5V rail inside the NUC is dirty. And really,
if it was that dirty, the PC should be "crash prone" as other
things in there could also be affected.

I don't think the problem is related to RAM, but it might be.
For example, if the RAM area used during CHKDSK was bad,
maybe that would cause some false analysis results. But then,
if that were true, the scan should have uncovered hundreds
of thousands of errors, and the cleanup via CHKDSK would
basically "shred" the drive. And the OS would not be able
to boot after that. So that's not it.

It's just a guess, but I'd have to say either the box is
near an ionizing radiation source, or the power is bad
at some level.

I've had "injected" noise in the computer room here. It's
possible for one ATX supply to place switching noise on the
AC. Devices sharing a common power strip, especially
sensitive devices (like my ADSL2 modem), they can be
upset by injected noise flowing right through the
power path and into sensitive analog circuits.

But if the +5V had that much ripple or spikes on it,
you'd think the machine would also be randomly crashing.
And you don't seem to be suggesting that in your various
posts.

It could be a bad Southbridge/PCH. The eMMC, I don't know
what bus that is on. Is it on SATA bus ? Is it on PCIe ?
I don't know how to figure that out, unless I could find
an eMMC datasheet.

OK, looking at an eMMC datasheet, it's a parallel data bus
running at 52MHz or so. That could be coming off a PCH or
SOC. Also, the eMMC runs off 3.3V and the SSD would be
running off 5V. It doesn't seem likely that both would
have exactly the same power problems. Or, that the analog
nature of SATA (low amplitude differential) would have
the same error characteristics as a full swing 3.3V digital bus
on an eMMC.

That leaves ionizing radiation. Or the low odds of
having bought *two* bad storage devices based on NAND
flash. Or the power conversion inside the NUC is a mess.

Replace the SSD (7mm) with a 7mm HDD ? If the SSD
was a regular SATA and not a microSATA, maybe that
would be a possible test case to run. See if even
a HDD has dirty results.

"Seagate ST1000LM048 128MB 5400 RPM 7mm $48"
https://www.newegg.com/Product/Produ...82E16822179108

If there was a source of ionizing radiation nearby, it
could also cause DRAM errors, and lots of running program
errors.

If you have a double-conversion UPS handy, try powering
the NUC off one of those. Double-conversion UPS don't
allow wall power to flow directly to the load. The power
is "filtered" through the UPS battery, and converted
to "fresh" AC via an inverter. You can get versions
of those with sinewave output (rather than stepped sine).

Paul

Overnight, Intel NUC PC (NU06CAY model [can't seem to find anything about it online]? from
2/2017 as shown in https://imgur.com/a/B0enW4H?) went to sleep when I was sleeping. I woke it
up and it again got stuck at its Intel NUC's black screen with its spinning animated cursor
again. Now, I can wait (going to do that since I need to do other stuff first) or force a power
off and on to force a chkdsk. I'll wait and then try a force power off and on later when I'm
more awake if it still stuck. It had been over 45 minutes now.


Paul wrote:
Ant wrote:
For a few days, I have been using the client's Intel NUC PC with its W10 b1803
weird issues. Originally, I was asked to look at why its W10 takes so long to
start/boot up. It seems to happen on each new day after power it on when it was off
for many hours like overnight. I can't seem to reproduce it after it does its scan
and repairs. I even tried changing BIOS' date to a day later to see if I could
reproduce it. Nope.

I wonder why it is always happen daily. Failing hardwares somewhere? In W10, I ran
W10's disk checks on both EMMC and SSD, and it found no issues. I know the issues
happened on both drives (original small EMMC and cloned bigger SSD). So, I decided
to force a chkdsk /r /f on both C: and D: drives outside of W10. C: was quick, but
D: took a while especially at 45% with intense light activities that long.

After returning to normal W10 session, I checked its Event Logs' Applications
section with its filter to show only 1001 event ID
(https://winaero.com/blog/how-to-find...n-windows-10/). I exported
and uploaded to http://zimage.com/~ant/temp/W10appEventLogs1001.txt. There were
lots of chkdsk repairs on both drives, crashes, etc.

What do you guys think is happening?

Is the computer in question near an ionizing radiation source ?

Uh, no? It happen in both my client's office and in my home's room.


What about power quality ? Where the NUC is plugged in, is
the power clean ? Could the problem be "injected" noise coming
from noisy machinery with inductive spikes ?

In the client's office, it is an office room. At my home, it is in my room with other computers
and electronics. I also have it connected to an APC UPS with my computers so that can't be it?


I see it is scanning a file system at rest. I could understand
scanning the C: drive, which is an active partition on each session,
and an inproper shutdown (drive isn't flushing properly), could
leave orphan this or that.

OK, so how do I find out and fix it?


But the distribution of errors is a bit weird. It doesn't
look entirely like dirty shutdowns. It's like storage
just isn't working properly.

Since you have a new SSD and it's being checked too, I'd have
to guess that the +5V rail inside the NUC is dirty. And really,
if it was that dirty, the PC should be "crash prone" as other
things in there could also be affected.

I don't think the problem is related to RAM, but it might be.
For example, if the RAM area used during CHKDSK was bad,
maybe that would cause some false analysis results. But then,
if that were true, the scan should have uncovered hundreds
of thousands of errors, and the cleanup via CHKDSK would
basically "shred" the drive. And the OS would not be able
to boot after that. So that's not it.

It's just a guess, but I'd have to say either the box is
near an ionizing radiation source, or the power is bad
at some level.

I've had "injected" noise in the computer room here. It's
possible for one ATX supply to place switching noise on the
AC. Devices sharing a common power strip, especially
sensitive devices (like my ADSL2 modem), they can be
upset by injected noise flowing right through the
power path and into sensitive analog circuits.

But if the +5V had that much ripple or spikes on it,
you'd think the machine would also be randomly crashing.
And you don't seem to be suggesting that in your various
posts.

It could be a bad Southbridge/PCH. The eMMC, I don't know
what bus that is on. Is it on SATA bus ? Is it on PCIe ?
I don't know how to figure that out, unless I could find
an eMMC datasheet.

OK, looking at an eMMC datasheet, it's a parallel data bus
running at 52MHz or so. That could be coming off a PCH or
SOC. Also, the eMMC runs off 3.3V and the SSD would be
running off 5V. It doesn't seem likely that both would
have exactly the same power problems. Or, that the analog
nature of SATA (low amplitude differential) would have
the same error characteristics as a full swing 3.3V digital bus
on an eMMC.

That leaves ionizing radiation. Or the low odds of
having bought *two* bad storage devices based on NAND
flash. Or the power conversion inside the NUC is a mess.

Replace the SSD (7mm) with a 7mm HDD ? If the SSD
was a regular SATA and not a microSATA, maybe that
would be a possible test case to run. See if even
a HDD has dirty results.

"Seagate ST1000LM048 128MB 5400 RPM 7mm $48"
https://www.newegg.com/Product/Produ...82E16822179108

If there was a source of ionizing radiation nearby, it
could also cause DRAM errors, and lots of running program
errors.

If you have a double-conversion UPS handy, try powering
the NUC off one of those. Double-conversion UPS don't
allow wall power to flow directly to the load. The power
is "filtered" through the UPS battery, and converted
to "fresh" AC via an inverter. You can get versions
of those with sinewave output (rather than stepped sine).

I have an APC UPS APC Back-UPS XS 1500 (model: BX1500G; 865 watts). No beeps from it.

This is sure difficult and annoying to figure out what's going on. All I can say this very
slow/hanging startup/boot up issue is still there since before assigned to work on this issue.
I am going to let the client know that this might be a hardware issue.

--
Quote of the Week: "As a thinker and planner, the ant is the equal of
any savage race of men; as a self-educated specialist in several arts
she is the superior of any savage race of men; and in one or two high
mental qualities she is above the reach of any man..." --Mark Twain
Note: A fixed width font (Courier, Monospace, etc.) is required to see this signature correctly.
/\___/\ Ant(Dude) @ http://aqfl.net & http://antfarm.home.dhs.org /
/ /\ /\ \ http://antfarm.ma.cx. Please nuke ANT if replying by e-mail.
| |o o| |
\ _ /
( )

I just came back and noticed its video was off (no signal) even though PC was on and had some
blue drive blinks. Huh?


Ant wrote:
Overnight, Intel NUC PC (NU06CAY model [can't seem to find anything about it online]? from
2/2017 as shown in https://imgur.com/a/B0enW4H?) went to sleep when I was sleeping. I woke it
up and it again got stuck at its Intel NUC's black screen with its spinning animated cursor
again. Now, I can wait (going to do that since I need to do other stuff first) or force a power
off and on to force a chkdsk. I'll wait and then try a force power off and on later when I'm
more awake if it still stuck. It had been over 45 minutes now.


Paul wrote:
Ant wrote:
For a few days, I have been using the client's Intel NUC PC with its W10 b1803
weird issues. Originally, I was asked to look at why its W10 takes so long to
start/boot up. It seems to happen on each new day after power it on when it was off
for many hours like overnight. I can't seem to reproduce it after it does its scan
and repairs. I even tried changing BIOS' date to a day later to see if I could
reproduce it. Nope.

I wonder why it is always happen daily. Failing hardwares somewhere? In W10, I ran
W10's disk checks on both EMMC and SSD, and it found no issues. I know the issues
happened on both drives (original small EMMC and cloned bigger SSD). So, I decided
to force a chkdsk /r /f on both C: and D: drives outside of W10. C: was quick, but
D: took a while especially at 45% with intense light activities that long.

After returning to normal W10 session, I checked its Event Logs' Applications
section with its filter to show only 1001 event ID
(https://winaero.com/blog/how-to-find...n-windows-10/). I exported
and uploaded to http://zimage.com/~ant/temp/W10appEventLogs1001.txt. There were
lots of chkdsk repairs on both drives, crashes, etc.

What do you guys think is happening?

Is the computer in question near an ionizing radiation source ?

Uh, no? It happen in both my client's office and in my home's room.


What about power quality ? Where the NUC is plugged in, is
the power clean ? Could the problem be "injected" noise coming
from noisy machinery with inductive spikes ?

In the client's office, it is an office room. At my home, it is in my room with other computers
and electronics. I also have it connected to an APC UPS with my computers so that can't be it?


I see it is scanning a file system at rest. I could understand
scanning the C: drive, which is an active partition on each session,
and an inproper shutdown (drive isn't flushing properly), could
leave orphan this or that.

OK, so how do I find out and fix it?


But the distribution of errors is a bit weird. It doesn't
look entirely like dirty shutdowns. It's like storage
just isn't working properly.

Since you have a new SSD and it's being checked too, I'd have
to guess that the +5V rail inside the NUC is dirty. And really,
if it was that dirty, the PC should be "crash prone" as other
things in there could also be affected.

I don't think the problem is related to RAM, but it might be.
For example, if the RAM area used during CHKDSK was bad,
maybe that would cause some false analysis results. But then,
if that were true, the scan should have uncovered hundreds
of thousands of errors, and the cleanup via CHKDSK would
basically "shred" the drive. And the OS would not be able
to boot after that. So that's not it.

It's just a guess, but I'd have to say either the box is
near an ionizing radiation source, or the power is bad
at some level.

I've had "injected" noise in the computer room here. It's
possible for one ATX supply to place switching noise on the
AC. Devices sharing a common power strip, especially
sensitive devices (like my ADSL2 modem), they can be
upset by injected noise flowing right through the
power path and into sensitive analog circuits.

But if the +5V had that much ripple or spikes on it,
you'd think the machine would also be randomly crashing.
And you don't seem to be suggesting that in your various
posts.

It could be a bad Southbridge/PCH. The eMMC, I don't know
what bus that is on. Is it on SATA bus ? Is it on PCIe ?
I don't know how to figure that out, unless I could find
an eMMC datasheet.

OK, looking at an eMMC datasheet, it's a parallel data bus
running at 52MHz or so. That could be coming off a PCH or
SOC. Also, the eMMC runs off 3.3V and the SSD would be
running off 5V. It doesn't seem likely that both would
have exactly the same power problems. Or, that the analog
nature of SATA (low amplitude differential) would have
the same error characteristics as a full swing 3.3V digital bus
on an eMMC.

That leaves ionizing radiation. Or the low odds of
having bought *two* bad storage devices based on NAND
flash. Or the power conversion inside the NUC is a mess.

Replace the SSD (7mm) with a 7mm HDD ? If the SSD
was a regular SATA and not a microSATA, maybe that
would be a possible test case to run. See if even
a HDD has dirty results.

"Seagate ST1000LM048 128MB 5400 RPM 7mm $48"
https://www.newegg.com/Product/Produ...82E16822179108

If there was a source of ionizing radiation nearby, it
could also cause DRAM errors, and lots of running program
errors.

If you have a double-conversion UPS handy, try powering
the NUC off one of those. Double-conversion UPS don't
allow wall power to flow directly to the load. The power
is "filtered" through the UPS battery, and converted
to "fresh" AC via an inverter. You can get versions
of those with sinewave output (rather than stepped sine).

I have an APC UPS APC Back-UPS XS 1500 (model: BX1500G; 865 watts). No beeps from it.

This is sure difficult and annoying to figure out what's going on. All I can say this very
slow/hanging startup/boot up issue is still there since before assigned to work on this issue.
I am going to let the client know that this might be a hardware issue.


--
Quote of the Week: "As a thinker and planner, the ant is the equal of
any savage race of men; as a self-educated specialist in several arts
she is the superior of any savage race of men; and in one or two high
mental qualities she is above the reach of any man..." --Mark Twain
Note: A fixed width font (Courier, Monospace, etc.) is required to see this signature correctly.
/\___/\ Ant(Dude) @ http://aqfl.net & http://antfarm.home.dhs.org /
/ /\ /\ \ http://antfarm.ma.cx. Please nuke ANT if replying by e-mail.
| |o o| |
\ _ /
( )

Still no video and disk activities seems to have stopped. Even its USB keyboard lights are not
toggling when I touch its caps lock, num lock, and scroll lock keys. I think it hard crashed
again. I will resume on it later. I need to do other non-IT stuff.


Ant wrote:
I just came back and noticed its video was off (no signal) even though PC was on and had some
blue drive blinks. Huh?


Ant wrote:
Overnight, Intel NUC PC (NU06CAY model [can't seem to find anything about it online]? from
2/2017 as shown in https://imgur.com/a/B0enW4H?) went to sleep when I was sleeping. I woke it
up and it again got stuck at its Intel NUC's black screen with its spinning animated cursor
again. Now, I can wait (going to do that since I need to do other stuff first) or force a power
off and on to force a chkdsk. I'll wait and then try a force power off and on later when I'm
more awake if it still stuck. It had been over 45 minutes now.


Paul wrote:
Ant wrote:
For a few days, I have been using the client's Intel NUC PC with its W10 b1803
weird issues. Originally, I was asked to look at why its W10 takes so long to
start/boot up. It seems to happen on each new day after power it on when it was off
for many hours like overnight. I can't seem to reproduce it after it does its scan
and repairs. I even tried changing BIOS' date to a day later to see if I could
reproduce it. Nope.

I wonder why it is always happen daily. Failing hardwares somewhere? In W10, I ran
W10's disk checks on both EMMC and SSD, and it found no issues. I know the issues
happened on both drives (original small EMMC and cloned bigger SSD). So, I decided
to force a chkdsk /r /f on both C: and D: drives outside of W10. C: was quick, but
D: took a while especially at 45% with intense light activities that long.

After returning to normal W10 session, I checked its Event Logs' Applications
section with its filter to show only 1001 event ID
(https://winaero.com/blog/how-to-find...n-windows-10/). I exported
and uploaded to http://zimage.com/~ant/temp/W10appEventLogs1001.txt. There were
lots of chkdsk repairs on both drives, crashes, etc.

What do you guys think is happening?

Is the computer in question near an ionizing radiation source ?

Uh, no? It happen in both my client's office and in my home's room.


What about power quality ? Where the NUC is plugged in, is
the power clean ? Could the problem be "injected" noise coming
from noisy machinery with inductive spikes ?

In the client's office, it is an office room. At my home, it is in my room with other computers
and electronics. I also have it connected to an APC UPS with my computers so that can't be it?


I see it is scanning a file system at rest. I could understand
scanning the C: drive, which is an active partition on each session,
and an inproper shutdown (drive isn't flushing properly), could
leave orphan this or that.

OK, so how do I find out and fix it?


But the distribution of errors is a bit weird. It doesn't
look entirely like dirty shutdowns. It's like storage
just isn't working properly.

Since you have a new SSD and it's being checked too, I'd have
to guess that the +5V rail inside the NUC is dirty. And really,
if it was that dirty, the PC should be "crash prone" as other
things in there could also be affected.

I don't think the problem is related to RAM, but it might be.
For example, if the RAM area used during CHKDSK was bad,
maybe that would cause some false analysis results. But then,
if that were true, the scan should have uncovered hundreds
of thousands of errors, and the cleanup via CHKDSK would
basically "shred" the drive. And the OS would not be able
to boot after that. So that's not it.

It's just a guess, but I'd have to say either the box is
near an ionizing radiation source, or the power is bad
at some level.

I've had "injected" noise in the computer room here. It's
possible for one ATX supply to place switching noise on the
AC. Devices sharing a common power strip, especially
sensitive devices (like my ADSL2 modem), they can be
upset by injected noise flowing right through the
power path and into sensitive analog circuits.

But if the +5V had that much ripple or spikes on it,
you'd think the machine would also be randomly crashing.
And you don't seem to be suggesting that in your various
posts.

It could be a bad Southbridge/PCH. The eMMC, I don't know
what bus that is on. Is it on SATA bus ? Is it on PCIe ?
I don't know how to figure that out, unless I could find
an eMMC datasheet.

OK, looking at an eMMC datasheet, it's a parallel data bus
running at 52MHz or so. That could be coming off a PCH or
SOC. Also, the eMMC runs off 3.3V and the SSD would be
running off 5V. It doesn't seem likely that both would
have exactly the same power problems. Or, that the analog
nature of SATA (low amplitude differential) would have
the same error characteristics as a full swing 3.3V digital bus
on an eMMC.

That leaves ionizing radiation. Or the low odds of
having bought *two* bad storage devices based on NAND
flash. Or the power conversion inside the NUC is a mess.

Replace the SSD (7mm) with a 7mm HDD ? If the SSD
was a regular SATA and not a microSATA, maybe that
would be a possible test case to run. See if even
a HDD has dirty results.

"Seagate ST1000LM048 128MB 5400 RPM 7mm $48"
https://www.newegg.com/Product/Produ...82E16822179108

If there was a source of ionizing radiation nearby, it
could also cause DRAM errors, and lots of running program
errors.

If you have a double-conversion UPS handy, try powering
the NUC off one of those. Double-conversion UPS don't
allow wall power to flow directly to the load. The power
is "filtered" through the UPS battery, and converted
to "fresh" AC via an inverter. You can get versions
of those with sinewave output (rather than stepped sine).

I have an APC UPS APC Back-UPS XS 1500 (model: BX1500G; 865 watts). No beeps from it.

This is sure difficult and annoying to figure out what's going on. All I can say this very
slow/hanging startup/boot up issue is still there since before assigned to work on this issue.
I am going to let the client know that this might be a hardware issue.



--
Quote of the Week: "As a thinker and planner, the ant is the equal of
any savage race of men; as a self-educated specialist in several arts
she is the superior of any savage race of men; and in one or two high
mental qualities she is above the reach of any man..." --Mark Twain
Note: A fixed width font (Courier, Monospace, etc.) is required to see this signature correctly.
/\___/\ Ant(Dude) @ http://aqfl.net & http://antfarm.home.dhs.org /
/ /\ /\ \ http://antfarm.ma.cx. Please nuke ANT if replying by e-mail.
| |o o| |
\ _ /
( )

I came back, and still got no video and input responses. I did notice its power light was
orange/amber instead of blue.
https://www.intel.com/content/www/us...-mini-pcs.html
seems to say this is sleep. I was forced to power off and on, run its scan and repair
automatically, and boot up fast again (for now).

I'm thinking of clean installing W10 into the 120 GB SSD even though I don't have the original
medias from client for his drivers, softwares, etc.


Ant wrote:
Still no video and disk activities seems to have stopped. Even its USB keyboard lights are not
toggling when I touch its caps lock, num lock, and scroll lock keys. I think it hard crashed
again. I will resume on it later. I need to do other non-IT stuff.


Ant wrote:
I just came back and noticed its video was off (no signal) even though PC was on and had some
blue drive blinks. Huh?


Ant wrote:
Overnight, Intel NUC PC (NU06CAY model [can't seem to find anything about it online]? from
2/2017 as shown in https://imgur.com/a/B0enW4H?) went to sleep when I was sleeping. I woke it
up and it again got stuck at its Intel NUC's black screen with its spinning animated cursor
again. Now, I can wait (going to do that since I need to do other stuff first) or force a power
off and on to force a chkdsk. I'll wait and then try a force power off and on later when I'm
more awake if it still stuck. It had been over 45 minutes now.


Paul wrote:
Ant wrote:
For a few days, I have been using the client's Intel NUC PC with its W10 b1803
weird issues. Originally, I was asked to look at why its W10 takes so long to
start/boot up. It seems to happen on each new day after power it on when it was off
for many hours like overnight. I can't seem to reproduce it after it does its scan
and repairs. I even tried changing BIOS' date to a day later to see if I could
reproduce it. Nope.

I wonder why it is always happen daily. Failing hardwares somewhere? In W10, I ran
W10's disk checks on both EMMC and SSD, and it found no issues. I know the issues
happened on both drives (original small EMMC and cloned bigger SSD). So, I decided
to force a chkdsk /r /f on both C: and D: drives outside of W10. C: was quick, but
D: took a while especially at 45% with intense light activities that long.

After returning to normal W10 session, I checked its Event Logs' Applications
section with its filter to show only 1001 event ID
(https://winaero.com/blog/how-to-find...n-windows-10/). I exported
and uploaded to http://zimage.com/~ant/temp/W10appEventLogs1001.txt. There were
lots of chkdsk repairs on both drives, crashes, etc.

What do you guys think is happening?

Is the computer in question near an ionizing radiation source ?

Uh, no? It happen in both my client's office and in my home's room.


What about power quality ? Where the NUC is plugged in, is
the power clean ? Could the problem be "injected" noise coming
from noisy machinery with inductive spikes ?

In the client's office, it is an office room. At my home, it is in my room with other computers
and electronics. I also have it connected to an APC UPS with my computers so that can't be it?


I see it is scanning a file system at rest. I could understand
scanning the C: drive, which is an active partition on each session,
and an inproper shutdown (drive isn't flushing properly), could
leave orphan this or that.

OK, so how do I find out and fix it?


But the distribution of errors is a bit weird. It doesn't
look entirely like dirty shutdowns. It's like storage
just isn't working properly.

Since you have a new SSD and it's being checked too, I'd have
to guess that the +5V rail inside the NUC is dirty. And really,
if it was that dirty, the PC should be "crash prone" as other
things in there could also be affected.

I don't think the problem is related to RAM, but it might be.
For example, if the RAM area used during CHKDSK was bad,
maybe that would cause some false analysis results. But then,
if that were true, the scan should have uncovered hundreds
of thousands of errors, and the cleanup via CHKDSK would
basically "shred" the drive. And the OS would not be able
to boot after that. So that's not it.

It's just a guess, but I'd have to say either the box is
near an ionizing radiation source, or the power is bad
at some level.

I've had "injected" noise in the computer room here. It's
possible for one ATX supply to place switching noise on the
AC. Devices sharing a common power strip, especially
sensitive devices (like my ADSL2 modem), they can be
upset by injected noise flowing right through the
power path and into sensitive analog circuits.

But if the +5V had that much ripple or spikes on it,
you'd think the machine would also be randomly crashing.
And you don't seem to be suggesting that in your various
posts.

It could be a bad Southbridge/PCH. The eMMC, I don't know
what bus that is on. Is it on SATA bus ? Is it on PCIe ?
I don't know how to figure that out, unless I could find
an eMMC datasheet.

OK, looking at an eMMC datasheet, it's a parallel data bus
running at 52MHz or so. That could be coming off a PCH or
SOC. Also, the eMMC runs off 3.3V and the SSD would be
running off 5V. It doesn't seem likely that both would
have exactly the same power problems. Or, that the analog
nature of SATA (low amplitude differential) would have
the same error characteristics as a full swing 3.3V digital bus
on an eMMC.

That leaves ionizing radiation. Or the low odds of
having bought *two* bad storage devices based on NAND
flash. Or the power conversion inside the NUC is a mess.

Replace the SSD (7mm) with a 7mm HDD ? If the SSD
was a regular SATA and not a microSATA, maybe that
would be a possible test case to run. See if even
a HDD has dirty results.

"Seagate ST1000LM048 128MB 5400 RPM 7mm $48"
https://www.newegg.com/Product/Produ...82E16822179108

If there was a source of ionizing radiation nearby, it
could also cause DRAM errors, and lots of running program
errors.

If you have a double-conversion UPS handy, try powering
the NUC off one of those. Double-conversion UPS don't
allow wall power to flow directly to the load. The power
is "filtered" through the UPS battery, and converted
to "fresh" AC via an inverter. You can get versions
of those with sinewave output (rather than stepped sine).

I have an APC UPS APC Back-UPS XS 1500 (model: BX1500G; 865 watts). No beeps from it.

This is sure difficult and annoying to figure out what's going on. All I can say this very
slow/hanging startup/boot up issue is still there since before assigned to work on this issue.
I am going to let the client know that this might be a hardware issue.




--
Quote of the Week: "As a thinker and planner, the ant is the equal of
any savage race of men; as a self-educated specialist in several arts
she is the superior of any savage race of men; and in one or two high
mental qualities she is above the reach of any man..." --Mark Twain
Note: A fixed width font (Courier, Monospace, etc.) is required to see this signature correctly.
/\___/\ Ant(Dude) @ http://aqfl.net & http://antfarm.home.dhs.org /
/ /\ /\ \ http://antfarm.ma.cx. Please nuke ANT if replying by e-mail.
| |o o| |
\ _ /
( )

On 21/02/2019 18:15, Ant wrote:

I'm thinking of clean installing W10 into the 120 GB SSD even though I don't have the original
medias from client for his drivers, softwares, etc.

How long does it take to think about this? Why not just do it and let
Windows installation finds the drivers for you. It is a clean install
so what have you got to lose? It takes not more that 30 minutes to do
this.

You guys need to be practical in life and don't waste time thinking
about things when it can be done almost immediately. Chinese & Indians
have taken over the technologies because they just do it and they also
started from base zero so they didn't have anything to lose!!!!!









--
With over 950 million devices now running Windows 10, customer
satisfaction is higher than any previous version of windows.

???? Good Guy ???? wrote:

On 21/02/2019 18:15, Ant wrote:

I'm thinking of clean installing W10 into the 120 GB SSD even though I don't have the original
medias from client for his drivers, softwares, etc.

How long does it take to think about this? Why not just do it and let
Windows installation finds the drivers for you. It is a clean install
so what have you got to lose? It takes not more that 30 minutes to do
this.
....
I just noticed my downloaded W10 v1809 installer file, through Media Creation Tool, got lost (0
byte) from overnight download. Ugh. I am going to assume chkdsk found it broken. Now, I need to
redownload that huge file again!

--
Quote of the Week: "As a thinker and planner, the ant is the equal of
any savage race of men; as a self-educated specialist in several arts
she is the superior of any savage race of men; and in one or two high
mental qualities she is above the reach of any man..." --Mark Twain
Note: A fixed width font (Courier, Monospace, etc.) is required to see this signature correctly.
/\___/\ Ant(Dude) @ http://aqfl.net & http://antfarm.home.dhs.org /
/ /\ /\ \ http://antfarm.ma.cx. Please nuke ANT if replying by e-mail.
| |o o| |
\ _ /
( )

Ant wrote:
???? Good Guy ???? wrote:

On 21/02/2019 18:15, Ant wrote:
I'm thinking of clean installing W10 into the 120 GB SSD even though I don't have the original
medias from client for his drivers, softwares, etc.

How long does it take to think about this? Why not just do it and let
Windows installation finds the drivers for you. It is a clean install
so what have you got to lose? It takes not more that 30 minutes to do
this.
...
I just noticed my downloaded W10 v1809 installer file, through Media Creation Tool, got lost (0
byte) from overnight download. Ugh. I am going to assume chkdsk found it broken. Now, I need to
redownload that huge file again!


Do that on your technician machine, not on the NUC.

You know your NUC has some sort of storage problem,
and neither downloads nor installs are going to be
worry free. If you want to make media, make it on
a reliable machine first, then visit the "victim"
and try your experiments.

Paul

Paul wrote:
Ant wrote:
???? Good Guy ???? wrote:

On 21/02/2019 18:15, Ant wrote:
I'm thinking of clean installing W10 into the 120 GB SSD even though I don't have the original
medias from client for his drivers, softwares, etc.

How long does it take to think about this? Why not just do it and let
Windows installation finds the drivers for you. It is a clean install
so what have you got to lose? It takes not more that 30 minutes to do
this.
...
I just noticed my downloaded W10 v1809 installer file, through Media Creation Tool, got lost (0
byte) from overnight download. Ugh. I am going to assume chkdsk found it broken. Now, I need to
redownload that huge file again!


Do that on your technician machine, not on the NUC.

You know your NUC has some sort of storage problem,
and neither downloads nor installs are going to be
worry free. If you want to make media, make it on
a reliable machine first, then visit the "victim"
and try your experiments.

OK. I hope it doesn't mess up my decade old, updated 64-bit W7 HPE SP1 PC. Maybe I can do it in my VM to avoid making a mess.

--
Quote of the Week: "As a thinker and planner, the ant is the equal of
any savage race of men; as a self-educated specialist in several arts
she is the superior of any savage race of men; and in one or two high
mental qualities she is above the reach of any man..." --Mark Twain
Note: A fixed width font (Courier, Monospace, etc.) is required to see this signature correctly.
/\___/\ Ant(Dude) @ http://aqfl.net & http://antfarm.home.dhs.org /
/ /\ /\ \ http://antfarm.ma.cx. Please nuke ANT if replying by e-mail.
| |o o| |
\ _ /
( )

Ant wrote:
Paul wrote:
Ant wrote:
???? Good Guy ???? wrote:

On 21/02/2019 18:15, Ant wrote:
I'm thinking of clean installing W10 into the 120 GB SSD even though I don't have the original
medias from client for his drivers, softwares, etc.
How long does it take to think about this? Why not just do it and let
Windows installation finds the drivers for you. It is a clean install
so what have you got to lose? It takes not more that 30 minutes to do
this.
...
I just noticed my downloaded W10 v1809 installer file, through Media Creation Tool, got lost (0
byte) from overnight download. Ugh. I am going to assume chkdsk found it broken. Now, I need to
redownload that huge file again!


Do that on your technician machine, not on the NUC.

You know your NUC has some sort of storage problem,
and neither downloads nor installs are going to be
worry free. If you want to make media, make it on
a reliable machine first, then visit the "victim"
and try your experiments.

OK. I hope it doesn't mess up my decade old, updated 64-bit W7 HPE SP1 PC. Maybe I can do it in my VM to avoid making a mess.


Note that the very most recent 1809 ISO is bigger than 4GB.
*Do not* download to a FAT32 partition :-)

(And the reason it is so big today, is because the
disc installs 12 different versions of Windows 10. Which
in my opinion is ridiculous. They should have put 8 versions
on one DVD and 4 related versions on a second DVD.)

Paul (who has tried to put a 4GB file on a FAT32, more than once...)

Paul wrote:
Ant wrote:
Paul wrote:
Ant wrote:
???? Good Guy ???? wrote:

On 21/02/2019 18:15, Ant wrote:
I'm thinking of clean installing W10 into the 120 GB SSD even though I don't have the original
medias from client for his drivers, softwares, etc.
How long does it take to think about this? Why not just do it and let
Windows installation finds the drivers for you. It is a clean install
so what have you got to lose? It takes not more that 30 minutes to do
this.
...
I just noticed my downloaded W10 v1809 installer file, through Media Creation Tool, got lost (0
byte) from overnight download. Ugh. I am going to assume chkdsk found it broken. Now, I need to
redownload that huge file again!


Do that on your technician machine, not on the NUC.

You know your NUC has some sort of storage problem,
and neither downloads nor installs are going to be
worry free. If you want to make media, make it on
a reliable machine first, then visit the "victim"
and try your experiments.

OK. I hope it doesn't mess up my decade old, updated 64-bit W7 HPE SP1 PC. Maybe I can do it in my VM to avoid making a mess.


Note that the very most recent 1809 ISO is bigger than 4GB.
*Do not* download to a FAT32 partition :-)

(And the reason it is so big today, is because the
disc installs 12 different versions of Windows 10. Which
in my opinion is ridiculous. They should have put 8 versions
on one DVD and 4 related versions on a second DVD.)

Paul (who has tried to put a 4GB file on a FAT32, more than once...)

Wow. I don't even have an external DVD drive so it has to be done on USB flash sticks or local drive.

--
Quote of the Week: "As a thinker and planner, the ant is the equal of
any savage race of men; as a self-educated specialist in several arts
she is the superior of any savage race of men; and in one or two high
mental qualities she is above the reach of any man..." --Mark Twain
Note: A fixed width font (Courier, Monospace, etc.) is required to see this signature correctly.
/\___/\ Ant(Dude) @ http://aqfl.net & http://antfarm.home.dhs.org /
/ /\ /\ \ http://antfarm.ma.cx. Please nuke ANT if replying by e-mail.
| |o o| |
\ _ /
( )

Ant wrote:
Paul wrote:
Ant wrote:
Paul wrote:
Ant wrote:
???? Good Guy ???? wrote:

On 21/02/2019 18:15, Ant wrote:
I'm thinking of clean installing W10 into the 120 GB SSD even though I don't have the original
medias from client for his drivers, softwares, etc.
How long does it take to think about this? Why not just do it and let
Windows installation finds the drivers for you. It is a clean install
so what have you got to lose? It takes not more that 30 minutes to do
this.
...
I just noticed my downloaded W10 v1809 installer file, through Media Creation Tool, got lost (0
byte) from overnight download. Ugh. I am going to assume chkdsk found it broken. Now, I need to
redownload that huge file again!

Do that on your technician machine, not on the NUC.
You know your NUC has some sort of storage problem,
and neither downloads nor installs are going to be
worry free. If you want to make media, make it on
a reliable machine first, then visit the "victim"
and try your experiments.
OK. I hope it doesn't mess up my decade old, updated 64-bit W7 HPE SP1 PC. Maybe I can do it in my VM to avoid making a mess.


Note that the very most recent 1809 ISO is bigger than 4GB.
*Do not* download to a FAT32 partition :-)

(And the reason it is so big today, is because the
disc installs 12 different versions of Windows 10. Which
in my opinion is ridiculous. They should have put 8 versions
on one DVD and 4 related versions on a second DVD.)

Paul (who has tried to put a 4GB file on a FAT32, more than once...)

Wow. I don't even have an external DVD drive so it has to be done on USB flash sticks or local drive.


For an Upgrade Install, you can just mount the ISO.

With the current ISO image being so big now, it no longer
fits on a single-layer DVD. You would need a dual layer
DVD, even if you had a USB DVD drive to use with the NUC.
I only have one blank left of those, and I'm not wasting
it on Windows. (There are several versions of 1809 media,
and if the size annoys you, you can use Heidoc to get one
which is 1GB smaller than the current one.)

But for USB keys, the image is still small enough that the program
I referred to in my other posting, will work with the 5GB image.
You can use this method for your USB key "Clean Install", where
the USB key will boot in either legacy or UEFI mode.

The first link is the README, the second link the download.

http://web.archive.org/web/201201022...usbdvd_dwnTool

http://web.archive.org/web/201110052...B-DVD-tool.exe

If you operate that tool from a 64-bit OS, it can make
32 bit or 64 bit USB sticks from
32 bit ISO or 64 bit ISO files.

If you want to do a UEFI install, the stick must be booted
in UEFI mode. This is easiest to see and select, by using
the "popup boot key" when booting the NUC. There will be
two entries for the USB key in the popup boot, and one entry
will have the word "UEFI" by it, as a key to selecting the
correct key boot method. When the NUC is POSTing, try pressing
the F10 key, as shown here.

https://01.org/projectceladon/sites/...ges/splash.jpg

The "Windows7-USB-DVD-tool.exe" will take the ISO and put
it on a USB stick for you. Example here.

https://i.postimg.cc/x1TddGrX/sample...rom-an-ISO.gif

And if that doesn't work, you can always try one of the other
ISO to USB key softwares. One problem with those, is they're
usually geared for Linux ISOs, rather than Windows.

Paul