From: William B. Lurie on
UNk and JJ:
Excuse me, but have I not been testing with all Norton
software not even loaded? Much less in service....
Everybody always suspects Norton, but how can they be
guilty in this case? What can I do to guarantee that
they are completely out of the picture?

Am I missing something here? My ground rules are
Clean Boot every time, load no more than it takes to make
the system run so that I can add services (and eventually
applications) and have a system that turns itself off as
it should after 2 hours of being idle.

Okay, John, I'm back to where the first batch of services
is on automatic, DCOM is manual and hibernate is okay. I'll
leave DCOM on manual, and see what running the next listed
service on automatic tells us.

John John - MVP wrote:
> My guess too, (Norton), and I've had it as a suspect from the very
> beginning.
>
> John
>
> Unknown wrote:
>> FYI--DCOM is set to automatic on my system and started. And I have no
>> problems with hibernate. I think
>> you are correct when you state something else interfacing/using DCOM
>> is the fault. My guess-----Norton simply
>> because of its notoriety. .
>> "John John - MVP" <audetweld(a)nbnot.nb.ca> wrote in message
>> news:uGc28rm1KHA.220(a)TK2MSFTNGP06.phx.gbl...
>>> William B. Lurie wrote:
>>>> William B. Lurie wrote:
>>>>> William B. Lurie wrote:
>>>>>> John John - MVP wrote:
>>>>>>> William B. Lurie wrote:
>>>>>>>> John John - MVP wrote:
>>>>>>>>> William B. Lurie wrote:
>>>>>>>>>> William B. Lurie wrote:
>>>>>>>>>> (snip)
>>>>>>>>>>>> You can set all the services to Manual start if you want but
>>>>>>>>>>>> I don't know if it will change anything, give it a try and
>>>>>>>>>>>> find out. Keep these three services to Automatic start:
>>>>>>>>>>>>
>>>>>>>>>>>> Event Log
>>>>>>>>>>>> Plug & Play
>>>>>>>>>>>> Remote Procedure Call (RPC)
>>>>>>>>>>>>
>>>>>>>>>>>> Set all the other services to manual or disabled. After you
>>>>>>>>>>>> do this try to hibernate the machine manually or in less
>>>>>>>>>>>> than one hour (to make sure that it can actually hibernate
>>>>>>>>>>>> with the minimal set of services).
>>>>>>>>>>>>
>>>>>>>>>>>> John
>>>>>>>>>>> Okay, John, will do....later.
>>>>>>>>>>> But for now, I'd like to report that my clone system is
>>>>>>>>>>> running, with clean boot, such that it does hibernate
>>>>>>>>>>> at 1 hour, giving no unexplainable event msgs or errors,
>>>>>>>>>>> and no sign of the dreaded ati2mtag or whatever it is.
>>>>>>>>>> I followed above procedures, John, with eminent success (so far).
>>>>>>>>>> Stripping down to just those few, it hibernated at two hours!
>>>>>>>>>> Did it twice to make sure.
>>>>>>>>>> So I changed the first ten to automatic and again, it
>>>>>>>>>> hibernated at
>>>>>>>>>> two hours. I did it twice to be sure.
>>>>>>>>> That is progress...
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>> So I changed the next 15 or so, and it did NOT go to 2 hours.
>>>>>>>>> And there is the clue to your problem. Keep notes of your
>>>>>>>>> changes and keep on narrowing things down. One of these 15
>>>>>>>>> services prevented hibernation, keep on whittling the list
>>>>>>>>> down, cut out 8 of the 15 services and see if things change.
>>>>>>>>> If it does hibernate move on to the other 7 services, if it
>>>>>>>>> doesn't hibernate cut the 8 services list down to 4 and try it
>>>>>>>>> again... and so on until you pinpoint it down to the culprit.
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>> Here are
>>>>>>>>>> the Events. Critical timing is that it started at 7:43, should
>>>>>>>>>> have
>>>>>>>>>> gone to 9:45, but things happened around 8:15, and desktop
>>>>>>>>>> came on
>>>>>>>>>> and I discontinued, to capture Events Monitor and pass them
>>>>>>>>>> along.
>>>>>>>>>> Perhaps you can interpret what they say, and tell me how to
>>>>>>>>>> fix it,
>>>>>>>>>> maybe by disabling some culprit......or do I search one
>>>>>>>>>> suspect at a
>>>>>>>>>> time until I find it? Got any clues? What is "Ci"?
>>>>>>>>>>
>>>>>>>>>>> Event Type: Information
>>>>>>>>>>> Event Source: Ci
>>>>>>>>>>> Event Category: CI Service Event ID: 4103
>>>>>>>>>>> Date: 4/6/2010
>>>>>>>>>>> Time: 8:16:59 AM
>>>>>>>>>>> User: N/A
>>>>>>>>>>> Computer: COMPAQ-2006
>>>>>>>>>>> Description:
>>>>>>>>>>> Master merge has completed on c:\system volume
>>>>>>>>>>> information\catalog.wci.
>>>>>>>>> That is caused by the Indexing Service. If you want you can
>>>>>>>>> read about the Master Merge in these search results:
>>>>>>>>>
>>>>>>>>> http://search.yahoo.com/search;_ylt=A0geu.JdQLtLUw0AQEWl87UF?p=MasterMergeTime&fr=yfp-t-501&fr2=sfp&iscqry=
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> You can disable the Indexing Service and see if things change.
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>> I can go back and narrow it down, but it would be nice to go
>>>>>>>>>> back to the
>>>>>>>>>> state where it ran 2 hours, and add just whatever you think is
>>>>>>>>>> the bad
>>>>>>>>>> actor.......
>>>>>>>>> I don't know who the bad actor is, you will have to keep on
>>>>>>>>> with the trial and error and find that bad actor! Remember
>>>>>>>>> what I said earlier... "sometimes you need dogged determination
>>>>>>>>> to get to the bottom of some of these problems".
>>>>>>>>>
>>>>>>>>> John
>>>>>>>> Right, John, I agree heartily. It will take some time, but I will
>>>>>>>> go back to what was good (first batch of changes from base state)
>>>>>>>> and start adding. Index will be the first that I re-enable after
>>>>>>>> I get back to a good 2 hour run. The ball is in my court; I will
>>>>>>>> crawl along toward what I hope will be a solution.
>>>>>>> Let us know what you find out!
>>>>>>>
>>>>>>> John
>>>>>> Now 10 OM. I did as I said I would. Started with stripped, bare
>>>>>> system.
>>>>>> Hibernation 2 hours okay.
>>>>>> I added Indexing *only*. It ran to 2 hours and hibernated perfectly.
>>>>>> No significant entries into events log.
>>>>>>
>>>>>> In a way this was good news......and in another way, bad news.
>>>>>> Obvious next step. Keep Indexing Automatic along with your 3 basic
>>>>>> automatics, and start adding one at a time, the stuff that ended up,
>>>>>> in combination, causing the malfunction. Tedious......actually
>>>>>> tedious squared, or tedious factorial, I'm not sure which.
>>>>>>
>>>>>> Your comments welcome, John, but the course seems pretty obvious.
>>>>> Follow-on next morning. I stretched too far. In addition to 5 basic
>>>>> plus Indexing, I added five more, and it blew again. And it isn't
>>>>> trivial to diagnose why, from the Log. Back to adding just one at a
>>>>> time. More later.
>>>> So the plot thickens. I went back to what worked, and re-activated
>>>> just one service, DCOM. It now refused to hibernate, and the event log
>>>> showed Service Control Manager error. I can go back and repeat the
>>>> steps if you feel it's not reasonable, John. But Now we are led from
>>>> DCOM (which is meaningless to me) to Service Control Manager, equally
>>>> meaningless, and I don't know if I need it in my running system, or if
>>>> I can safely leave it Manual (or Disabled) and go on with the next
>>>> service. Next step will be to make DCOM Manual again and wait for
>>>> advice.
>>> I don't think that DCOM in itself is responsible, more likely it's
>>> another application than is using DCOM that would be at fault. On a
>>> production machine this service needs to be set to Automatic Start.
>>> What errors are you seeing in the Event Log?
>>>
>>> For your trouble shooting purposes you could leave DCOM to manual for
>>> the time being and keep on with your other necessary services and see
>>> what happens.
>>>
>>> John
>>
>>
From: William B. Lurie on
William B. Lurie wrote:
> UNk and JJ:
> Excuse me, but have I not been testing with all Norton
> software not even loaded? Much less in service....
> Everybody always suspects Norton, but how can they be
> guilty in this case? What can I do to guarantee that
> they are completely out of the picture?
>
> Am I missing something here? My ground rules are
> Clean Boot every time, load no more than it takes to make
> the system run so that I can add services (and eventually
> applications) and have a system that turns itself off as
> it should after 2 hours of being idle.
>
> Okay, John, I'm back to where the first batch of services
> is on automatic, DCOM is manual and hibernate is okay. I'll
> leave DCOM on manual, and see what running the next listed
> service on automatic tells us.
>
> John John - MVP wrote:
>> My guess too, (Norton), and I've had it as a suspect from the very
>> beginning.
>>
>> John
>>
>> Unknown wrote:
>>> FYI--DCOM is set to automatic on my system and started. And I have no
>>> problems with hibernate. I think
>>> you are correct when you state something else interfacing/using DCOM
>>> is the fault. My guess-----Norton simply
>>> because of its notoriety. .
>>> "John John - MVP" <audetweld(a)nbnot.nb.ca> wrote in message
>>> news:uGc28rm1KHA.220(a)TK2MSFTNGP06.phx.gbl...
>>>> William B. Lurie wrote:
>>>>> William B. Lurie wrote:
>>>>>> William B. Lurie wrote:
>>>>>>> John John - MVP wrote:
>>>>>>>> William B. Lurie wrote:
>>>>>>>>> John John - MVP wrote:
>>>>>>>>>> William B. Lurie wrote:
>>>>>>>>>>> William B. Lurie wrote:
>>>>>>>>>>> (snip)
>>>>>>>>>>>>> You can set all the services to Manual start if you want
>>>>>>>>>>>>> but I don't know if it will change anything, give it a try
>>>>>>>>>>>>> and find out. Keep these three services to Automatic start:
>>>>>>>>>>>>>
>>>>>>>>>>>>> Event Log
>>>>>>>>>>>>> Plug & Play
>>>>>>>>>>>>> Remote Procedure Call (RPC)
>>>>>>>>>>>>>
>>>>>>>>>>>>> Set all the other services to manual or disabled. After
>>>>>>>>>>>>> you do this try to hibernate the machine manually or in
>>>>>>>>>>>>> less than one hour (to make sure that it can actually
>>>>>>>>>>>>> hibernate with the minimal set of services).
>>>>>>>>>>>>>
>>>>>>>>>>>>> John
>>>>>>>>>>>> Okay, John, will do....later.
>>>>>>>>>>>> But for now, I'd like to report that my clone system is
>>>>>>>>>>>> running, with clean boot, such that it does hibernate
>>>>>>>>>>>> at 1 hour, giving no unexplainable event msgs or errors,
>>>>>>>>>>>> and no sign of the dreaded ati2mtag or whatever it is.
>>>>>>>>>>> I followed above procedures, John, with eminent success (so
>>>>>>>>>>> far).
>>>>>>>>>>> Stripping down to just those few, it hibernated at two hours!
>>>>>>>>>>> Did it twice to make sure.
>>>>>>>>>>> So I changed the first ten to automatic and again, it
>>>>>>>>>>> hibernated at
>>>>>>>>>>> two hours. I did it twice to be sure.
>>>>>>>>>> That is progress...
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>> So I changed the next 15 or so, and it did NOT go to 2 hours.
>>>>>>>>>> And there is the clue to your problem. Keep notes of your
>>>>>>>>>> changes and keep on narrowing things down. One of these 15
>>>>>>>>>> services prevented hibernation, keep on whittling the list
>>>>>>>>>> down, cut out 8 of the 15 services and see if things change.
>>>>>>>>>> If it does hibernate move on to the other 7 services, if it
>>>>>>>>>> doesn't hibernate cut the 8 services list down to 4 and try it
>>>>>>>>>> again... and so on until you pinpoint it down to the culprit.
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>> Here are
>>>>>>>>>>> the Events. Critical timing is that it started at 7:43,
>>>>>>>>>>> should have
>>>>>>>>>>> gone to 9:45, but things happened around 8:15, and desktop
>>>>>>>>>>> came on
>>>>>>>>>>> and I discontinued, to capture Events Monitor and pass them
>>>>>>>>>>> along.
>>>>>>>>>>> Perhaps you can interpret what they say, and tell me how to
>>>>>>>>>>> fix it,
>>>>>>>>>>> maybe by disabling some culprit......or do I search one
>>>>>>>>>>> suspect at a
>>>>>>>>>>> time until I find it? Got any clues? What is "Ci"?
>>>>>>>>>>>
>>>>>>>>>>>> Event Type: Information
>>>>>>>>>>>> Event Source: Ci
>>>>>>>>>>>> Event Category: CI Service Event ID: 4103
>>>>>>>>>>>> Date: 4/6/2010
>>>>>>>>>>>> Time: 8:16:59 AM
>>>>>>>>>>>> User: N/A
>>>>>>>>>>>> Computer: COMPAQ-2006
>>>>>>>>>>>> Description:
>>>>>>>>>>>> Master merge has completed on c:\system volume
>>>>>>>>>>>> information\catalog.wci.
>>>>>>>>>> That is caused by the Indexing Service. If you want you can
>>>>>>>>>> read about the Master Merge in these search results:
>>>>>>>>>>
>>>>>>>>>> http://search.yahoo.com/search;_ylt=A0geu.JdQLtLUw0AQEWl87UF?p=MasterMergeTime&fr=yfp-t-501&fr2=sfp&iscqry=
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> You can disable the Indexing Service and see if things change.
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>> I can go back and narrow it down, but it would be nice to go
>>>>>>>>>>> back to the
>>>>>>>>>>> state where it ran 2 hours, and add just whatever you think
>>>>>>>>>>> is the bad
>>>>>>>>>>> actor.......
>>>>>>>>>> I don't know who the bad actor is, you will have to keep on
>>>>>>>>>> with the trial and error and find that bad actor! Remember
>>>>>>>>>> what I said earlier... "sometimes you need dogged
>>>>>>>>>> determination to get to the bottom of some of these problems".
>>>>>>>>>>
>>>>>>>>>> John
>>>>>>>>> Right, John, I agree heartily. It will take some time, but I will
>>>>>>>>> go back to what was good (first batch of changes from base state)
>>>>>>>>> and start adding. Index will be the first that I re-enable after
>>>>>>>>> I get back to a good 2 hour run. The ball is in my court; I will
>>>>>>>>> crawl along toward what I hope will be a solution.
>>>>>>>> Let us know what you find out!
>>>>>>>>
>>>>>>>> John
>>>>>>> Now 10 OM. I did as I said I would. Started with stripped, bare
>>>>>>> system.
>>>>>>> Hibernation 2 hours okay.
>>>>>>> I added Indexing *only*. It ran to 2 hours and hibernated perfectly.
>>>>>>> No significant entries into events log.
>>>>>>>
>>>>>>> In a way this was good news......and in another way, bad news.
>>>>>>> Obvious next step. Keep Indexing Automatic along with your 3 basic
>>>>>>> automatics, and start adding one at a time, the stuff that ended up,
>>>>>>> in combination, causing the malfunction. Tedious......actually
>>>>>>> tedious squared, or tedious factorial, I'm not sure which.
>>>>>>>
>>>>>>> Your comments welcome, John, but the course seems pretty obvious.
>>>>>> Follow-on next morning. I stretched too far. In addition to 5 basic
>>>>>> plus Indexing, I added five more, and it blew again. And it isn't
>>>>>> trivial to diagnose why, from the Log. Back to adding just one at
>>>>>> a time. More later.
>>>>> So the plot thickens. I went back to what worked, and re-activated
>>>>> just one service, DCOM. It now refused to hibernate, and the event log
>>>>> showed Service Control Manager error. I can go back and repeat the
>>>>> steps if you feel it's not reasonable, John. But Now we are led
>>>>> from DCOM (which is meaningless to me) to Service Control Manager,
>>>>> equally
>>>>> meaningless, and I don't know if I need it in my running system, or if
>>>>> I can safely leave it Manual (or Disabled) and go on with the next
>>>>> service. Next step will be to make DCOM Manual again and wait for
>>>>> advice.
>>>> I don't think that DCOM in itself is responsible, more likely it's
>>>> another application than is using DCOM that would be at fault. On a
>>>> production machine this service needs to be set to Automatic Start.
>>>> What errors are you seeing in the Event Log?
>>>>
>>>> For your trouble shooting purposes you could leave DCOM to manual
>>>> for the time being and keep on with your other necessary services
>>>> and see what happens.
>>>>
>>>> John
>>>
>>>
John, the next test was inconclusive because I may not have
waited quite 2 hours, but it is already beginning to look as
though the length of time to test with DCOM plus each succeeding
service, and then the next one plus each of the rest, will end
some time in the next decade, not this one. I'll pursue this tack
a bit further, but it seems to me that *somebody*, maybe another
one of the excellent MVPs, might know which 'service' has a one
hour time clock built into it.
From: John John - MVP on
William B. Lurie wrote:
> William B. Lurie wrote:
>> UNk and JJ:
>> Excuse me, but have I not been testing with all Norton
>> software not even loaded? Much less in service....
>> Everybody always suspects Norton, but how can they be
>> guilty in this case? What can I do to guarantee that
>> they are completely out of the picture?
>>
>> Am I missing something here? My ground rules are
>> Clean Boot every time, load no more than it takes to make
>> the system run so that I can add services (and eventually
>> applications) and have a system that turns itself off as
>> it should after 2 hours of being idle.
>>
>> Okay, John, I'm back to where the first batch of services
>> is on automatic, DCOM is manual and hibernate is okay. I'll
>> leave DCOM on manual, and see what running the next listed
>> service on automatic tells us.
>>
>> John John - MVP wrote:
>>> My guess too, (Norton), and I've had it as a suspect from the very
>>> beginning.
>>>
>>> John
>>>
>>> Unknown wrote:
>>>> FYI--DCOM is set to automatic on my system and started. And I have
>>>> no problems with hibernate. I think
>>>> you are correct when you state something else interfacing/using DCOM
>>>> is the fault. My guess-----Norton simply
>>>> because of its notoriety. .
>>>> "John John - MVP" <audetweld(a)nbnot.nb.ca> wrote in message
>>>> news:uGc28rm1KHA.220(a)TK2MSFTNGP06.phx.gbl...
>>>>> William B. Lurie wrote:
>>>>>> William B. Lurie wrote:
>>>>>>> William B. Lurie wrote:
>>>>>>>> John John - MVP wrote:
>>>>>>>>> William B. Lurie wrote:
>>>>>>>>>> John John - MVP wrote:
>>>>>>>>>>> William B. Lurie wrote:
>>>>>>>>>>>> William B. Lurie wrote:
>>>>>>>>>>>> (snip)
>>>>>>>>>>>>>> You can set all the services to Manual start if you want
>>>>>>>>>>>>>> but I don't know if it will change anything, give it a try
>>>>>>>>>>>>>> and find out. Keep these three services to Automatic start:
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Event Log
>>>>>>>>>>>>>> Plug & Play
>>>>>>>>>>>>>> Remote Procedure Call (RPC)
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Set all the other services to manual or disabled. After
>>>>>>>>>>>>>> you do this try to hibernate the machine manually or in
>>>>>>>>>>>>>> less than one hour (to make sure that it can actually
>>>>>>>>>>>>>> hibernate with the minimal set of services).
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> John
>>>>>>>>>>>>> Okay, John, will do....later.
>>>>>>>>>>>>> But for now, I'd like to report that my clone system is
>>>>>>>>>>>>> running, with clean boot, such that it does hibernate
>>>>>>>>>>>>> at 1 hour, giving no unexplainable event msgs or errors,
>>>>>>>>>>>>> and no sign of the dreaded ati2mtag or whatever it is.
>>>>>>>>>>>> I followed above procedures, John, with eminent success (so
>>>>>>>>>>>> far).
>>>>>>>>>>>> Stripping down to just those few, it hibernated at two hours!
>>>>>>>>>>>> Did it twice to make sure.
>>>>>>>>>>>> So I changed the first ten to automatic and again, it
>>>>>>>>>>>> hibernated at
>>>>>>>>>>>> two hours. I did it twice to be sure.
>>>>>>>>>>> That is progress...
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>> So I changed the next 15 or so, and it did NOT go to 2 hours.
>>>>>>>>>>> And there is the clue to your problem. Keep notes of your
>>>>>>>>>>> changes and keep on narrowing things down. One of these 15
>>>>>>>>>>> services prevented hibernation, keep on whittling the list
>>>>>>>>>>> down, cut out 8 of the 15 services and see if things change.
>>>>>>>>>>> If it does hibernate move on to the other 7 services, if it
>>>>>>>>>>> doesn't hibernate cut the 8 services list down to 4 and try
>>>>>>>>>>> it again... and so on until you pinpoint it down to the culprit.
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>> Here are
>>>>>>>>>>>> the Events. Critical timing is that it started at 7:43,
>>>>>>>>>>>> should have
>>>>>>>>>>>> gone to 9:45, but things happened around 8:15, and desktop
>>>>>>>>>>>> came on
>>>>>>>>>>>> and I discontinued, to capture Events Monitor and pass them
>>>>>>>>>>>> along.
>>>>>>>>>>>> Perhaps you can interpret what they say, and tell me how to
>>>>>>>>>>>> fix it,
>>>>>>>>>>>> maybe by disabling some culprit......or do I search one
>>>>>>>>>>>> suspect at a
>>>>>>>>>>>> time until I find it? Got any clues? What is "Ci"?
>>>>>>>>>>>>
>>>>>>>>>>>>> Event Type: Information
>>>>>>>>>>>>> Event Source: Ci
>>>>>>>>>>>>> Event Category: CI Service Event ID: 4103
>>>>>>>>>>>>> Date: 4/6/2010
>>>>>>>>>>>>> Time: 8:16:59 AM
>>>>>>>>>>>>> User: N/A
>>>>>>>>>>>>> Computer: COMPAQ-2006
>>>>>>>>>>>>> Description:
>>>>>>>>>>>>> Master merge has completed on c:\system volume
>>>>>>>>>>>>> information\catalog.wci.
>>>>>>>>>>> That is caused by the Indexing Service. If you want you can
>>>>>>>>>>> read about the Master Merge in these search results:
>>>>>>>>>>>
>>>>>>>>>>> http://search.yahoo.com/search;_ylt=A0geu.JdQLtLUw0AQEWl87UF?p=MasterMergeTime&fr=yfp-t-501&fr2=sfp&iscqry=
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> You can disable the Indexing Service and see if things change.
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>> I can go back and narrow it down, but it would be nice to go
>>>>>>>>>>>> back to the
>>>>>>>>>>>> state where it ran 2 hours, and add just whatever you think
>>>>>>>>>>>> is the bad
>>>>>>>>>>>> actor.......
>>>>>>>>>>> I don't know who the bad actor is, you will have to keep on
>>>>>>>>>>> with the trial and error and find that bad actor! Remember
>>>>>>>>>>> what I said earlier... "sometimes you need dogged
>>>>>>>>>>> determination to get to the bottom of some of these problems".
>>>>>>>>>>>
>>>>>>>>>>> John
>>>>>>>>>> Right, John, I agree heartily. It will take some time, but I will
>>>>>>>>>> go back to what was good (first batch of changes from base state)
>>>>>>>>>> and start adding. Index will be the first that I re-enable after
>>>>>>>>>> I get back to a good 2 hour run. The ball is in my court; I will
>>>>>>>>>> crawl along toward what I hope will be a solution.
>>>>>>>>> Let us know what you find out!
>>>>>>>>>
>>>>>>>>> John
>>>>>>>> Now 10 OM. I did as I said I would. Started with stripped, bare
>>>>>>>> system.
>>>>>>>> Hibernation 2 hours okay.
>>>>>>>> I added Indexing *only*. It ran to 2 hours and hibernated
>>>>>>>> perfectly.
>>>>>>>> No significant entries into events log.
>>>>>>>>
>>>>>>>> In a way this was good news......and in another way, bad news.
>>>>>>>> Obvious next step. Keep Indexing Automatic along with your 3 basic
>>>>>>>> automatics, and start adding one at a time, the stuff that ended
>>>>>>>> up,
>>>>>>>> in combination, causing the malfunction. Tedious......actually
>>>>>>>> tedious squared, or tedious factorial, I'm not sure which.
>>>>>>>>
>>>>>>>> Your comments welcome, John, but the course seems pretty obvious.
>>>>>>> Follow-on next morning. I stretched too far. In addition to 5 basic
>>>>>>> plus Indexing, I added five more, and it blew again. And it isn't
>>>>>>> trivial to diagnose why, from the Log. Back to adding just one at
>>>>>>> a time. More later.
>>>>>> So the plot thickens. I went back to what worked, and re-activated
>>>>>> just one service, DCOM. It now refused to hibernate, and the event
>>>>>> log
>>>>>> showed Service Control Manager error. I can go back and repeat the
>>>>>> steps if you feel it's not reasonable, John. But Now we are led
>>>>>> from DCOM (which is meaningless to me) to Service Control Manager,
>>>>>> equally
>>>>>> meaningless, and I don't know if I need it in my running system,
>>>>>> or if
>>>>>> I can safely leave it Manual (or Disabled) and go on with the next
>>>>>> service. Next step will be to make DCOM Manual again and wait for
>>>>>> advice.
>>>>> I don't think that DCOM in itself is responsible, more likely it's
>>>>> another application than is using DCOM that would be at fault. On
>>>>> a production machine this service needs to be set to Automatic
>>>>> Start. What errors are you seeing in the Event Log?
>>>>>
>>>>> For your trouble shooting purposes you could leave DCOM to manual
>>>>> for the time being and keep on with your other necessary services
>>>>> and see what happens.
>>>>>
>>>>> John
>>>>
>>>>
> John, the next test was inconclusive because I may not have
> waited quite 2 hours, but it is already beginning to look as
> though the length of time to test with DCOM plus each succeeding
> service, and then the next one plus each of the rest, will end
> some time in the next decade, not this one. I'll pursue this tack
> a bit further, but it seems to me that *somebody*, maybe another
> one of the excellent MVPs, might know which 'service' has a one
> hour time clock built into it.

I would enable them in batches of 10 or more services, not one at a
time, this should speed things up a bit. Also, I *always* keep these
services disabled on almost any machine that I ever work with:

Alerter
Clipbook
Human Interface Device Access
Net.Tcp Port Sharing Service
Network DDE
Network DDE DSDM
Remote Registry
SSDP Discovery Service
Telnet
Universal Plug and Play Device Host

I suggest you set them to disabled and forget about them, so that's a
batch of 10 services out of the way!

You can also use Process Monitor and see what it captures, it will show
you which processes were running and at what time they ran.

http://technet.microsoft.com/en-us/sysinternals/bb896645.aspx

John
From: William B. Lurie on


>>>>
> John, the next test was inconclusive because I may not have
> waited quite 2 hours, but it is already beginning to look as
> though the length of time to test with DCOM plus each succeeding
> service, and then the next one plus each of the rest, will end
> some time in the next decade, not this one. I'll pursue this tack
> a bit further, but it seems to me that *somebody*, maybe another
> one of the excellent MVPs, might know which 'service' has a one
> hour time clock built into it.

I've snipped everything out because I think we're at a decision-
making point. I just don't think that the try-each-service approach
is going to be practical. Just too many services, too many possible
combinations, too little known about each, too long to make just one
test. I agree it's logical and sensible and proper scientific
technique, but it's like counting the grains of sand on the beach.

Getting back to the 1 hour/2 hour problem, I feel we have eliminated
any running application program as a source, by just not loading them.
If they don't load, they don't execute, and if they don't execute,
they can't influence hibernation.

So what *is* running? The system and its big-brother-given 'services'.
Rereading my last comment above, I hope I didn't offend anybody, but we
have to realize how enormously complex the XP system is, and it's too
much to expect any MVP to be intimately familiar with the inner workings
of all of its services. That's approaching the problem from the bottom
up. The question, down from the top, is, can we pick enough MVPs' brains
hard enough to find out which of the services is capable of preventing
the supposedly idle system from hibernating after 2 hours, but not after
only 1 hour.
From: John John - MVP on
William B. Lurie wrote:
>
>
>>>>>
>> John, the next test was inconclusive because I may not have
>> waited quite 2 hours, but it is already beginning to look as
>> though the length of time to test with DCOM plus each succeeding
>> service, and then the next one plus each of the rest, will end
>> some time in the next decade, not this one. I'll pursue this tack
>> a bit further, but it seems to me that *somebody*, maybe another
>> one of the excellent MVPs, might know which 'service' has a one
>> hour time clock built into it.
>
> I've snipped everything out because I think we're at a decision-
> making point. I just don't think that the try-each-service approach
> is going to be practical.

No, it isn't, that is why it's best to do it in batches of 10 services
or so...


> [snip...}

> So what *is* running?

Sysinternals' Process Monitor will tell you that...

John