Vista - Scheduler wakes, doesn't wait for task to end, then goes back to s

I have a task to run an executable. I have it set to run whether the user is
logged in or not, run with the highest privileges, wake the computer to run
this task, stop the task if it runs longer than 8 hours. It runs at 1am. The
task wakes the computer up at 1am fine, starts the executable, then puts the
computer back to sleep without waiting for the executable (a VS2005 SP1 C#
program I created) to complete. When I wake the computer in the morning, the
executable resumes.

What is up with this? Why can't the task scheduler wait until the task it
spawned finishes? My vista installation was an upgrade from XP Pro. But I
created the task (several times) from scratch in Vista. I also recompiled the
executable in Vista.

I have also tried the task in both "Windows Server 2003, Windows XP, or
Windows 2000" mode and in "Windows Vista" mode.

This is a huuuuuge step backward for me. I would appreciate any help anyone
can offer for this.

On Sat, 17 Mar 2007 07:10:10 -0700, Pete Furtado <Pete
Furtado@discussions.microsoft.com> wrote:

>I have a task to run an executable. I have it set to run whether the user is
>logged in or not, run with the highest privileges, wake the computer to run
>this task, stop the task if it runs longer than 8 hours. It runs at 1am. The
>task wakes the computer up at 1am fine, starts the executable, then puts the
>computer back to sleep without waiting for the executable (a VS2005 SP1 C#
>program I created) to complete. When I wake the computer in the morning, the
>executable resumes.

>What is up with this? Why can't the task scheduler wait until the task it
>spawned finishes? My vista installation was an upgrade from XP Pro. But I
>created the task (several times) from scratch in Vista. I also recompiled the
>executable in Vista.

Is the PC set to "suspend" after some time of inactivity?

If so, turn that off - as that may cut in just because there's no
keyboard or mouse activity going on.

Are there any settings within the Task's properties to control this
sort of behavior?



>--------------- ---- --- -- - - - -
Saws are too hard to use.
Be easier to use!
>--------------- ---- --- -- - - - -

Thanks for the reply.

It is set to go into standby after 2 hours. But it seems to go right back to
sleep after spawning the task. My executable writes informational messages to
the event log and I see it writing messages for about 30 seconds to a minute,
then nothing. I'm assuming that is when the machine goes back into standby.

I will double-check this tonight, though, and change the power scheme to
just leave the machine on all night and see what happens.

I can't find any other settings in the task's properties other than what I
stated in my original post. If there's some "hidden" ones, I'm dying to find
out.

I'll update this thread tomorrow with the results.

Thanks again.

"cquirke (MVP Windows shell/user)" wrote:

> On Sat, 17 Mar 2007 07:10:10 -0700, Pete Furtado <Pete
> Furtado@discussions.microsoft.com> wrote:
>
> >I have a task to run an executable. I have it set to run whether the user is
> >logged in or not, run with the highest privileges, wake the computer to run
> >this task, stop the task if it runs longer than 8 hours. It runs at 1am. The
> >task wakes the computer up at 1am fine, starts the executable, then puts the
> >computer back to sleep without waiting for the executable (a VS2005 SP1 C#
> >program I created) to complete. When I wake the computer in the morning, the
> >executable resumes.
>
> >What is up with this? Why can't the task scheduler wait until the task it
> >spawned finishes? My vista installation was an upgrade from XP Pro. But I
> >created the task (several times) from scratch in Vista. I also recompiled the
> >executable in Vista.
>
> Is the PC set to "suspend" after some time of inactivity?
>
> If so, turn that off - as that may cut in just because there's no
> keyboard or mouse activity going on.
>
> Are there any settings within the Task's properties to control this
> sort of behavior?
>
>
>
> >--------------- ---- --- -- - - - -
> Saws are too hard to use.
> Be easier to use!
> >--------------- ---- --- -- - - - -
>

"Adds" <adds21@yahoo.com> wrote in message
news:eOy6ReVbHHA.1508@TK2MSFTNGP06.phx.gbl...
> Humm... It's certainly annoying.

I think I might have found away around this, although not ideal.

In my batch file, I'm running powercfg first to set the PC into "High
Performance" mode (powercfg /? will give the options). I then set it back
to my normal power mode after the batch file has complete.

I have to admit that I haven't given it long enough to test properly, but it
appears to work.

Not ideal I know, but at least it might be a work around.

A.

"Adds" wrote:

>
> "Adds" <adds21@yahoo.com> wrote in message
> news:eOy6ReVbHHA.1508@TK2MSFTNGP06.phx.gbl...
> > Humm... It's certainly annoying.
>
> I think I might have found away around this, although not ideal.
>
> In my batch file, I'm running powercfg first to set the PC into "High
> Performance" mode (powercfg /? will give the options). I then set it back
> to my normal power mode after the batch file has complete.
>
> I have to admit that I haven't given it long enough to test properly, but it
> appears to work.
>
> Not ideal I know, but at least it might be a work around.
>
> A.
>
>
>

Interesting idea. I got so disgusted with this problem that I gave up for a
while.

Anyway, how has it worked for you? I will try this myself in the next couple
of days. Thanks!
Pete

I've got two Vista machines doing this as well. Using a batch file to switch
to high performance mode where sleep is set to "never" before the task runs
doesn't fix the problem. Both machines drop back to sleep within a minute or
two of task start regardless of power mode. I originally ran across this
problem with defrag and thought it might have to do with the fact that defrag
does low-priority (background) I/O, but I've reproduced the problem with
several different disk I/O intensive tasks that use normal priority I/O so
that's not it (and just in case anybody is wondering: removing the -i switch
on defrag has no affect either).

Looks like a Vista power management bug to me. Windows has certainly had
it's fair share of these over the years. Lets hope the bug gets fixed in SP1
or (hopefully) sooner in an update.

Clayton Gautier

"Pete Furtado" wrote:

>
>
> "Adds" wrote:
>
> >
> > "Adds" <adds21@yahoo.com> wrote in message
> > news:eOy6ReVbHHA.1508@TK2MSFTNGP06.phx.gbl...
> > > Humm... It's certainly annoying.
> >
> > I think I might have found away around this, although not ideal.
> >
> > In my batch file, I'm running powercfg first to set the PC into "High
> > Performance" mode (powercfg /? will give the options). I then set it back
> > to my normal power mode after the batch file has complete.
> >
> > I have to admit that I haven't given it long enough to test properly, but it
> > appears to work.
> >
> > Not ideal I know, but at least it might be a work around.
> >
> > A.
> >
> >
> >
>
> Interesting idea. I got so disgusted with this problem that I gave up for a
> while.
>
> Anyway, how has it worked for you? I will try this myself in the next couple
> of days. Thanks!
> Pete
>

In case anyone still cares about this issue, I've come up with a work-around.
It involves the Win32 API function SetThreadExecutionState (kernel32.dll).
You can find the details on the web, but basically, Windows can be forced to
stay awake while a process runs by sandwiching the process between two
SetThreadExecutionState calls (with different arguments for the calls). The
sandwiching can be implemented in a variety of ways - I chose to create an
executable, named StayAwakeExec in VB 2005 Express Edition (free) that does
the following:

1) makes the first call to SetThreadExecutionState to block sleep
2) spawns an executable passed in as a command line argument
3) makes the second call to SetThreadExecutionState to allow sleep after
the spawned executable exits

Note that the "after the spawned executable exits" part of Step 3 is
implemented by calling Process.WaitForExit() for the spawned executable.

I've successfully used StayAwakeExec to spawn defrag.exe with a task that
wakes the computer at 1AM. The computer now wakes, runs defrag to
completion, and then goes back to sleep immediately after the defrag finishes
(as verified by the system log file).

I was hoping that the latest Vista performance and reliability patches
(KB938194 and KB938979) would fix the bug, design flaw - whatever you want to
call it - that made this workaround necessary, but alas, twas not so. I
tested both my systems after installing these patches and Vista still falls
asleep almost exactly 2 minutes after a scheduled task wakes the computer
regardless of whether the task is done. My expectation (and I'm pretty sure
most people would agree) is if you check the "Wake the computer to run this
task" checkbox on the Conditions tab in Task Schedular, Windows should stay
awake until the task is complete. It should be a simple matter for an MS
developer to add to the task schedular service the sandwiching calls to
SetThreadExecutionState to prevent sleep if "Wake the computer to run this
task" is checked. Hopefully SP1 will have this fix.


"clayga" wrote:

> I've got two Vista machines doing this as well. Using a batch file to switch
> to high performance mode where sleep is set to "never" before the task runs
> doesn't fix the problem. Both machines drop back to sleep within a minute or
> two of task start regardless of power mode. I originally ran across this
> problem with defrag and thought it might have to do with the fact that defrag
> does low-priority (background) I/O, but I've reproduced the problem with
> several different disk I/O intensive tasks that use normal priority I/O so
> that's not it (and just in case anybody is wondering: removing the -i switch
> on defrag has no affect either).
>
> Looks like a Vista power management bug to me. Windows has certainly had
> it's fair share of these over the years. Lets hope the bug gets fixed in SP1
> or (hopefully) sooner in an update.
>
> Clayton Gautier
>
> "Pete Furtado" wrote:
>
> >
> >
> > "Adds" wrote:
> >
> > >
> > > "Adds" <adds21@yahoo.com> wrote in message
> > > news:eOy6ReVbHHA.1508@TK2MSFTNGP06.phx.gbl...
> > > > Humm... It's certainly annoying.
> > >
> > > I think I might have found away around this, although not ideal.
> > >
> > > In my batch file, I'm running powercfg first to set the PC into "High
> > > Performance" mode (powercfg /? will give the options). I then set it back
> > > to my normal power mode after the batch file has complete.
> > >
> > > I have to admit that I haven't given it long enough to test properly, but it
> > > appears to work.
> > >
> > > Not ideal I know, but at least it might be a work around.
> > >
> > > A.
> > >
> > >
> > >
> >
> > Interesting idea. I got so disgusted with this problem that I gave up for a
> > while.
> >
> > Anyway, how has it worked for you? I will try this myself in the next couple
> > of days. Thanks!
> > Pete
> >

Thanks for the tip, clayga! That works most of the time on my pc. I did
basically the same thing, i have a separate executable calling
SetThreadExecutionState, and then I just call that first from vbscript, do
anything else I need to do via vbscript, then kill that task and exit the
script.

Sorry for the late reply. I couldn't log into this site for a while for some
reason.

I still have more problems with this stupid Vista task scheduler. Sometimes
tasks just don't run. Yes, yes, i know all the options to wake computer to
run this task, and run whether user is logged on or not. They just did a
really, really bad implementation of the task scheduler in Vista. I hope they
fix it one of these days.

Thanks again!

Pete Furtado
"clayga" wrote:

Having the same problem with Vista Ultimate 32 bit, but that vb script stuff is way over my head...