Hibernation after delay when no timeout set.

[Grant Robertson]

I have a tablet PC but I don't think this problem is specifically related
to that fact so I have cross-posted into the general and tabletpc groups.

My laptop goes into hibernation after a short period of not being used
(about 10 minutes) even though in my Power Schemes tab all the settings
are set to "Never." The really weird part is that this only happens after
I have resumed from hibernation (regardless of battery level). This
problem never occurs after a clean boot.

In other words: If I do a clean boot and then let my laptop set for hours
without touching it, then it will never go into hibernation (unless I
manually put it into hibernation, of course). However, if I put my laptop
into hibernation and then resume, while I am using it - if I don't touch
it for about 10 minutes - then it will go into hibernation.

Again, "Enable hibernation" is selected on the "Hibernation" tab of the
"Power Options Properties" dialog. But all settings on the "Power
Schemes" tabs are set to never (and properly saved).


I don't need any answers saying, "If you are not going to touch it for 10
minutes then what do you care if it goes into hibernation." I use my
tablet PC to take notes in class and quite often go over 10 minutes
without needing to write something down. During those times I will
manually blank the screen to save power (but it is not set to do that
automatically). When I do need to take notes, I need to write them right
now, not after resuming from hibernation. What is worse is when I get
ready to write something down and the thing has just started going into
hibernation. I have to wait for it to hibernate then wait for it to
resume before I can write something into my notes. By then the teacher
has moved well on and I have missed something.

This problem essentially makes the hibernation feature useless for me. If
I hibernate the thing then I have to reboot it immediately after resuming
from hibernation if I don't want it to keep dropping into hibernation
when I don't want it to. Heck, if I was going to do that then I might as
well have just shut the thing all the way down in the first place.

In the end, I am left using Standby mode instead. That works OK and is
faster to wake up but sometimes I want to be able to hibernate it so I
can save my place with all my open apps but not run down the battery if I
forget the thing in my bag overnight.

 

[Beverly Howard [Ms-MVP/MobileDev]]

Take a look in the screen saver options and assure that it's not there

Beverly Howard [MS MVP-Mobile Devices]

 

[Invalid]

I have a tablet PC but I don't think this problem is specifically related
to that fact so I have cross-posted into the general and tabletpc groups.

My laptop goes into hibernation after a short period of not being used
(about 10 minutes) even though in my Power Schemes tab all the settings
are set to "Never." The really weird part is that this only happens after
I have resumed from hibernation (regardless of battery level). This
problem never occurs after a clean boot.

In other words: If I do a clean boot and then let my laptop set for hours
without touching it, then it will never go into hibernation (unless I
manually put it into hibernation, of course). However, if I put my laptop
into hibernation and then resume, while I am using it - if I don't touch
it for about 10 minutes - then it will go into hibernation.

Grant,

I have had similar issues with my Tablet going into Hibernate of its own
volition. I can't nail down the cause or the sequence of actions which
cause it (but I have my suspicions - see below). It seems to be caused
by it picking up the wrong power scheme.

What has worked for me is to go into the Registry

HKCU\Control Panel\PowerCfg\PowerPolicies\

There are a set of numbered subkeys

Each time you modify and save a new power scheme it seems to create a
new numbered scheme. Modifying and re-saving a scheme creates another
numbered key rather than modifying the existing key (this may only be
when the scheme being modified is the one in use - I haven't really
experimented). Neither does it seem to stop you saving a scheme with a
name that already exists. But it doesn't seem to replace the old one.

So I occasionally end up with more than one scheme with the same name in
the set, or the numbered order mixed up. Either seems to confuse the
system totally.

My fix is to Apply one of the original schemes (I use Always On). Delete
all the schemes except the original set using regedit, reboot then go
back and use the power config applet to recreate the scheme I want.

The usual caveats apply to mucking about with the registry .

Regards

 

[Grant Robertson]

What has worked for me is to go into the Registry ..
..
..
My fix is to Apply one of the original schemes (I use Always On). Delete
all the schemes except the original set using regedit, reboot then go
back and use the power config applet to recreate the scheme I want.

Thanks, I will definitely give that a try. I do recall it haveing a hard
time saving my power scheme under the name "Always On" so that may be the
issue.

 

[Grant Robertson]

Guess what? I went into the registry, deleted all of the schemes except
"Always On" and then renamed "Always On" to "Always On2" and renamed the
key to "0". When I went back to the "Power Schemes" dialog the only one
available was, naturally, "Always On2." But get this, the Hibernation was
set to 15 minutes for on battery.

It is as if the real setting was 15 minutes but Windows was showing Never
in the dialog instead. This is the lamest bug I have ever seen.

Now get this... Every tme I rename the Key to "0" it resets the on
battery hibernation timeout to 15 minutes.

....So I just tested this fix. It didn't make any difference. Not only
that, but it still goes into hibernation after less than 10 minutes, not
15.

I guess I am just going to have to accept the fact that windows is full
of bugs. Not just vulnerabilities. But simply lame bugs. If MS would put
half the effort into fixing bugs that they put into developing new stuff
that nobody cares about, then maybe Linux wouldn't be gaining so much
ground.

On top of everything else, now regedit won't open. I get the 0xc0000005
error. But only when I type regedit in lower case and not when I type it
in upper case. Yet another lame bug. That's four I have discovered in
just this past 45 minutes.

If I wasn't so hooked on the Tablet PC and didn't need to learn Adobe
products then I would switch to Linux for sure.

 

[Bob I]

There is a design feature that if you come out of hibernation, and then
you have 5 minutes of no activity it will go back into hibernation.

 

[Invalid]

There is a design feature that if you come out of hibernation, and then
you have 5 minutes of no activity it will go back into hibernation.

Thanks for that Bob,

That's something that has bugged me for a few years. I run an old Laptop
as a fileserver and webserver (mainly for testing purposes). If I shut
it down at night by hibernating it, then the following morning when
restarted, unless I enter the password and do something it goes back
into hibernation DESPITE being always on.

I hate people who install defaults that can't be modified.

 

[Josh Einstein]

Grant, no offense but I'm sure you did something to cause this. I've never
heard of this issue until now. I know lots of people tweak services and
registry entries and I'm sure you're like some of us other power users and
have done the same. But I get annoyed when I hear people blame Vista for
being "buggy" when it's probably something us power users have done to cause
it.

You're tweaking around in the registry now and I'm sure you're pretty
comfortable with it so I'm sure you've done it before. Don't take this as an
insult. I've screwed things up too (specifically, with Windows indexer) but
this is not a commonly reported problem so please don't assume it's just
Vista bugginess. And yes I am on a crusade these days. I am going to start
demanding that everyone who claims a Vista bug prove that it is in fact a
bug. If it affects just you then it's 99 times out of 100 not a bug.

If you can prove bugs or vulnerabilities, then please do so. But please do
not surmise based on your individual experience. When it comes to
vulnerabilities and bug reports there is a process, please follow it.

Can you reproduce it in a virtual machine?

"No More Mister Nice Guy"
- Josh

 

[Bob I]

Thanks for that Bob,

That's something that has bugged me for a few years. I run an old Laptop
as a fileserver and webserver (mainly for testing purposes). If I shut
it down at night by hibernating it, then the following morning when
restarted, unless I enter the password and do something it goes back
into hibernation DESPITE being always on.

I hate people who install defaults that can't be modified.

Found the article about why.

Returning from hibernation sets the inactivity timer to five minutes on
a Windows XP-based computer
http://support.microsoft.com/kb/318355/en-us

 

[Kocureq]

Josh Einstein pisze:

Can you reproduce it in a virtual machine?

No need for that, just read Microsoft's own words:

http://support.microsoft.com/kb/318355/en-us

Please don't tell users that they're a cause for problems that are there
BY DESIGN. And, of course, cannot be turned off, so every morning I have
to click something on my Tablet PC to make it work, download mail,
instant messaging etc. when I get my coffea. Otherwise, it will just go
to sleep again.

Thank you microsoft. But you don't have to fix this "feature" - you've
got zealots that will defend you and tell the users it's their fault.

 

[Bob I]

Josh Einstein pisze:



No need for that, just read Microsoft's own words:

http://support.microsoft.com/kb/318355/en-us

Please don't tell users that they're a cause for problems that are there
BY DESIGN. And, of course, cannot be turned off, so every morning I have
to click something on my Tablet PC to make it work, download mail,
instant messaging etc. when I get my coffea. Otherwise, it will just go
to sleep again.

Thank you microsoft. But you don't have to fix this "feature" - you've
got zealots that will defend you and tell the users it's their fault.

A PC that goes to sleep if you don't use it, what a novel idea!

 

[Josh Einstein]

Maybe you should re-read the post. His problem is that it's returning to
hibernation *after he has used it* after resuming. That is not what the KB
article talks about. It talks about the situation where Windows wakes from
hibernation, has no activity, then hibernates again which is by design and
is intended to protect you from opening your backpack and finding a hot dead
battery. Why would they "fix" that?

And Grant's a frequent poster around these parts and I know he's not the
average "log on to check my mail" user. After all, he thinks Linux is
gaining ground on Windows

But anyway, when people post messages with a chip on their shoulder and act
like they know the scoop (Windows being full of "lame bugs" and security
vulnerabilities) when clearly they're just guessing, I'm gonna be a little
cranky in my response.

So now maybe we can try to find a real explanation for this instead of just
bitching back and forth.

For example, maybe the pen usage isn't being seen by Windows as activity?
Seems unlikely cause you would think that other people would have reported
their systems ignoring their standby settings, but I'm playing devil's
advocate. Bottom line is none of my machines do this and I've never seen it
mentioned before so it's a bit premature to start calling it one of the lame
bugs that Vista is full of.

--
Josh Einstein (Tablet PC MVP)
Einstein Technologies
Tablet Enhancements for Outlook - Try it free: www.tabletoutlook.com

Josh Einstein pisze:

Can you reproduce it in a virtual machine?

No need for that, just read Microsoft's own words:

http://support.microsoft.com/kb/318355/en-us

Please don't tell users that they're a cause for problems that are there
BY DESIGN. And, of course, cannot be turned off, so every morning I have
to click something on my Tablet PC to make it work, download mail, instant
messaging etc. when I get my coffea. Otherwise, it will just go to sleep
again.

Thank you microsoft. But you don't have to fix this "feature" - you've got
zealots that will defend you and tell the users it's their fault.

--
/\ /\ [ Jakub 'Kocureq' Anderwald ] /\ /\
=^;^= [ [nick][at][nick].com ] =^;^=
/ | [ GG# 1365999 ICQ# 31547220 ] | \
(___(|_|_| [ (e-mail address removed) ] |_|_|)___)

 

[Invalid]

A PC that goes to sleep if you don't use it, what a novel idea!

Bob,

The issue here is that Microsoft's apparent definition of "use" is ONLY
keyboard or mouse activity. Using the PC as a server that has network
file activity doesn't (apparently) count as use.

Personally I regard accessing a PC over the network these days as
"using" it, and would expect the machine to stay awake and not go back
into hibernation.

How do you react when your media centre server goes back to sleep after
5 minutes in the middle of serving a film or music track because the
only use the O/S has seen is your accessing video or MP3 files over the
network, or did they fix this "feature" for that version.

From a "green" perspective I see no reason why I should have to choose
between leaving the machine running 7/24 or shutting down completely and
restarting all the applications from cold when I reboot.

Why should I not be able to hibernate my servers when I have finished
using them for the day, and revive them the next time I need them.

 

[Invalid]

Maybe you should re-read the post. His problem is that it's returning
to hibernation *after he has used it* after resuming. That is not what
the KB article talks about. It talks about the situation where Windows
wakes from hibernation, has no activity, then hibernates again which is
by design and is intended to protect you from opening your backpack and
finding a hot dead battery. Why would they "fix" that?

And Grant's a frequent poster around these parts and I know he's not
the average "log on to check my mail" user. After all, he thinks Linux
is gaining ground on Windows

But anyway, when people post messages with a chip on their shoulder and
act like they know the scoop (Windows being full of "lame bugs" and
security vulnerabilities) when clearly they're just guessing, I'm gonna
be a little cranky in my response.

So now maybe we can try to find a real explanation for this instead of
just bitching back and forth.

For example, maybe the pen usage isn't being seen by Windows as
activity? Seems unlikely cause you would think that other people would
have reported their systems ignoring their standby settings, but I'm
playing devil's advocate. Bottom line is none of my machines do this
and I've never seen it mentioned before so it's a bit premature to
start calling it one of the lame bugs that Vista is full of.

Josh,

Good post.

I HAVE seen this problem, and as I suggested to Grant, I think it's a
problem with the way windows XP handles user customised power schemes.

I think Grant has gone over the top in deleting ALL the power schemes
except one. In my original post I suggested he delete
"...all the schemes except the original set.."

As I said I have had problems implementing custom power schemes. This
has come down to trying various combinations of power settings. The
process of saving them seems to cause the confusion.

Once I settled on a particular scheme then cleaned out the registry to
leave only the core set plus my particular user setting, I have had no
further trouble.

In Grant's case I would suspect that the core power settings do things
that are not accessible via the user interface, trying to use one of the
core power scheme numbers (0) as a user defined scheme may have caused
his issues.

 

[Bob I]

Bob,

The issue here is that Microsoft's apparent definition of "use" is ONLY
keyboard or mouse activity. Using the PC as a server that has network
file activity doesn't (apparently) count as use.

Personally I regard accessing a PC over the network these days as
"using" it, and would expect the machine to stay awake and not go back
into hibernation.

How do you react when your media centre server goes back to sleep after
5 minutes in the middle of serving a film or music track because the
only use the O/S has seen is your accessing video or MP3 files over the
network, or did they fix this "feature" for that version.

From a "green" perspective I see no reason why I should have to choose
between leaving the machine running 7/24 or shutting down completely and
restarting all the applications from cold when I reboot.

Why should I not be able to hibernate my servers when I have finished
using them for the day, and revive them the next time I need them.

Since in this case you are trying to use it in a manner that it wasn't
designed for, you will need to do a little more work to pull it off. A
simple means would be to disable hibernate when you revive it and
re-enable when you want to put it to sleep.

 

[Josh Einstein]

Well a server shouldn't be configured to sleep and an application such as
media center should use the API's that keep a PC from entering sleep or
standby as a result of user inactivity.

How would you like to open your backpack and find the PC woke up and didn't
go to sleep because of all the bluetooth activity in the air? I think it's
reasonable to say that if there's no mouse or keyboard activity and the
computer is configured to sleep, it should do so. Otherwise anything could
be considered activity. Disk activity? Accelerometer activity?

 

[Josh Einstein]

Actually you are correct, there is quite a bit of stuff in the power schemes
that are not user-accessible. That is why you can't create one from scratch
but can only clone one and edit it. There are various utilities to edit this
info but it's not for the faint of heart. It'd be worth a night of reading
TechNet articles to understand what they do before tweaking them. But one of
the questions I asked never got answered and that was whether there were any
OEM-installed power management utilities.

My Asus R2H came with a utility called Power4Gear thad had its own power
profile, but more importantly, instead of changing the selected profile, it
modified values within that profile when the user selects different
operational modes. This worked ok on XP but maybe not on Vista. I am always
leary about using these third party power management utilities because I
question the ability of a third party to better manage the power situation
than the OS and power management drivers.

 

[Josh Einstein]

After re-reading your post more carefully I'll address another point you
made about why can't we power servers "on demand". It's a nice idea, but
it's currently not how the hardware and software is designed. If the market
demands it, it'll happen.

But with regard to this problem, it only affects hibernate and most PC's
support magic packets over LAN to wake the machine. So you can potentially
get huge power savings by using sleep instead of hibernate and wake on lan.
According to the KB article, only hibernate (which is an OS feature, the
hardware really thinks it's just off) clears the information in BIOS that
windows uses to determine why the machine woke. So presumably, the WOL
feature and sleep should work ok.

But either way, it's still off topic but that's just my thoughts on that.

 

[Invalid]

Well a server shouldn't be configured to sleep and an application such
as media center should use the API's that keep a PC from entering sleep
or standby as a result of user inactivity.

How would you like to open your backpack and find the PC woke up and
didn't go to sleep because of all the bluetooth activity in the air? I
think it's reasonable to say that if there's no mouse or keyboard
activity and the computer is configured to sleep, it should do so.
Otherwise anything could be considered activity. Disk activity?
Accelerometer activity?

Josh,

My issue is not that such behaviour happens - I agree that as a default
it is very useful and desirable.

Its that there appears to be no way to alter this particular default.

Setting a machine's power configuration to ALWAYS ON means (In my
humble opinion) that the machine should stay on when switched on.

Not

It will stay on after coming out of hibernation only after you have used
the keyboard or mouse, but if you don't it will switch itself off
whether you like it or not.

I wouldn't like to open my case and find the machine had flattened its
own battery - but then I wouldn't configure a machine I carried around
as "always on".

I agree (as per your second post) that sleep does not have the same
problem - but sleep (on XP & W2K - I don't know about Vista) uses a lot
more power over the year than hibernate (the CPU and memory stay
powered, albeit at a low level). A simple way to stop the thing going
back into hibernation (based on (a - non default - power setting) would
allow me to fire up servers in the office when needed without going
there.

Interestingly I don't get the same behaviour on wake from hibernation
on a desktop box. It's only machines configured with batteries!! So
again my fault for reusing obsolete laptops rather than buying new.

 

[Invalid]

So now maybe we can try to find a real explanation for this instead of
just bitching back and forth.

I have been experimenting and I think I know what is happening and
possibly how it can be worked around. This is on an XP tablet. Someone
else needs to check Vista.

Symptoms
I have a user defined power policy I use for reading PDF's saved (using
Save As) in my power settings. It is set to stop the disks after 2
minutes blank the screen after 5 Minutes, standby in 15 and hibernate
after 30 Minutes.

After a cold boot it does exactly that. After recovering from
hibernation it goes straight back into hibernation after 5 minutes if I
don't use it at all, or 5 minutes after I have stopped using it with no
prior screen blanking or standby period.

The 5 minute hibernate is the timer described in
http://support.microsoft.com/kb/318355/en-us.

The drop straight into hibernate after some real use+5 minutes idle is
the problem, as this is not the desired behaviour set in the power
policy.

Analysis
XP on recovery from hibernation is picking up the wrong power policy.
Saved user defined policies are assigned a number in sequence following
on from the default (0-5). The number of the policy in use is stored in
[HKEY_CURRENT_USER\Control Panel\PowerCfg] "CurrentPowerPolicy"="n"
where n is the number of the policy in use.

On recovery from Hibernation XP picks up the correct value of "n", but
uses it to apply the "n" power policy from the default user
[HKEY_USERS\.DEFAULT\Control Panel\PowerCfg]

Not the "n" policy from the current user
[HKEY_CURRENT_USER\Control Panel\PowerCfg]

This is not a real problem if the user is using one of the six standard
policies - it just seems to go back to the defaults which I suspect most
people leave alone anyway. The problem arises when the user has saved a
particular power policy, so n does not exist in the Default User list -
then the default five minute hibernate timer seems to kick back in.

[I have my machine set up to request a password on recovery from
hibernate - that MAY be the reason it goes to the default user. I think
it is also a problem in W2K. On W2k I seem to remember recovery from
Hibernation picking up the default user desktop colour scheme rather
than the current users until you entered the password].

Fix
The workaround is to ensure that the user's saved policy is defined in
the Default user power config list with the same number "n".

I achieved this by exporting the user saved policy (in my case numbered
6) to a .reg file, editing the .reg file to change occurrences of

HKEY_CURRENT_USER to HKEY_USERS\.DEFAULT

and importing that file back into the registry so that my policy appears
in the default user power configuration list with the correct value of
"n".

On my tablet that fixes the problem, after recovery from hibernation it
now behaves (as far as I can tell from a couple of tries) in line with
the power configuration scheme.

I hope this helps someone. There IS a bug there somewhere.

PS for Gary
One way to fix the mess in your registry is to export the whole
[HKEY_USERS\.DEFAULT\Control Panel\PowerCfg] key set.
This contains all the standard defaults.
Edit it the other way i.e. change all occurrences of

HKEY_USERS\.DEFAULT to HKEY_CURRENT_USER

Then re-import, that should get you back to a clean power config set.
Reboot. Create your own. Export it, edit it and re-import it into the
default user as above. Let us know how you get on.

A good idea might then be to dump the whole current user set as a
backup.

Thank you for asking the question, you have caused me to fix something
that has been frustrating me for at least a year, but I couldn't figure
out what was happening!!!

Regards