After a fresh boot I am sitting around 1.55g used. I can load Firefox
and it gets up around 1.7. As the day goes on usage goes up, just
browsing to over 6 spiking to the top of the box (8) with no windows
open but I don't see any processes using that much. I am guessing a
script has it's thumb on memory it isn't using. How do I find this
memory hog?

On 2/24/2024 1:07 AM, gfretwell@aol.com wrote:
> After a fresh boot I am sitting around 1.55g used. I can load Firefox
> and it gets up around 1.7. As the day goes on usage goes up, just
> browsing to over 6 spiking to the top of the box (8) with no windows
> open but I don't see any processes using that much. I am guessing a
> script has it's thumb on memory it isn't using. How do I find this
> memory hog?
>

"Pushing the Limits of Windows: Paged and Nonpaged Pool"

https://web.archive.org/web/20151109174823/http://blogs.technet.com:80/b/markrussinovich/archive/2009/03/26/3211216.aspx

Another example.

https://techcommunity.microsoft.com/t5/core-infrastructure-and-security/another-troubleshooting-adventure-more-real-life-memory-pool/ba-p/257050

One thing that might leak, is Windows Update and friends. You might find
that just leaving the system running, with no Firefox at all, the memory
consumption is going up.

And I have seen a system get to the level of displaying "Delayed Write Fail",
and that's because an attempt to write to disk, didn't come back for
more than ten seconds or so. And that's because the system was furiously
scrambling to find some memory to allow the write to happen. There was
a leaker, running at the time, so this wasn't an unknown source.

Paul

<gfretwell@aol.com> wrote

I am guessing a
| script has it's thumb on memory it isn't using. How do I find this
| memory hog?

Process Explorer. "Working set" seems to be the closest thing
to actual RAM usage. Alsio perhaps check FF cache. I set it to
about 10 MB because cache is rarely used these days. In the
56K days browsers would cache webpages for revisits. Today,
nearly every webpage is beiing generated "new", so there are
few 304 server responses. And with fast speeds thheres no need
to cache small files.

There are numerous other things. Unnecessary services? AV that's
checking everything that moves? Also, IE brings down all earlier
systems by tying into Explorer. I'm not sure about Win7. But you can
set temp internet files to something like 10 MB and clear the cache,
anyway.

On 2/24/2024 8:20 AM, Newyana2 wrote:
> <gfretwell@aol.com> wrote
>
> I am guessing a
> | script has it's thumb on memory it isn't using. How do I find this
> | memory hog?
>
> Process Explorer. "Working set" seems to be the closest thing
> to actual RAM usage. Alsio perhaps check FF cache. I set it to
> about 10 MB because cache is rarely used these days. In the
> 56K days browsers would cache webpages for revisits. Today,
> nearly every webpage is beiing generated "new", so there are
> few 304 server responses. And with fast speeds thheres no need
> to cache small files.
>
> There are numerous other things. Unnecessary services? AV that's
> checking everything that moves? Also, IE brings down all earlier
> systems by tying into Explorer. I'm not sure about Win7. But you can
> set temp internet files to something like 10 MB and clear the cache,
> anyway.

Some browsers offer an option to have a file cache on disk (default)
or in RAM. So it could be one or the other.

The cache still gets used. Because I sometimes clean it out
before a backup (as part of general removal of cruft).

As another example, before a backup you can remove the contents of LCU,
which could contain 100,000 files. And can be deleted. LCU stands
for Last Cumulative Update (it's for an Undo function on Windows Update).

Firefox has had runaway memory consumption. But when Firefox exits,
that RAM isn't leaked and it is returned to the operating system properly.

Things that leak, could be leaking at kernel level (down where the
drivers live). Paged or unpaged pool. The problem for a Win7 user,
is finding versions of debug tools that work on Win7 (because Microsoft
removes stuff like that).

Paul

On 2/24/2024 8:22 AM, Paul wrote:
> On 2/24/2024 8:20 AM, Newyana2 wrote:
>> <gfretwell@aol.com> wrote
>>
>> I am guessing a
>> | script has it's thumb on memory it isn't using. How do I find this
>> | memory hog?
>>
>> Process Explorer. "Working set" seems to be the closest thing
>> to actual RAM usage. Alsio perhaps check FF cache. I set it to
>> about 10 MB because cache is rarely used these days. In the
>> 56K days browsers would cache webpages for revisits. Today,
>> nearly every webpage is beiing generated "new", so there are
>> few 304 server responses. And with fast speeds thheres no need
>> to cache small files.
>>
>> There are numerous other things. Unnecessary services? AV that's
>> checking everything that moves? Also, IE brings down all earlier
>> systems by tying into Explorer. I'm not sure about Win7. But you can
>> set temp internet files to something like 10 MB and clear the cache,
>> anyway.
>
> Some browsers offer an option to have a file cache on disk (default)
> or in RAM. So it could be one or the other.
>
> The cache still gets used. Because I sometimes clean it out
> before a backup (as part of general removal of cruft).
>
> As another example, before a backup you can remove the contents of LCU,
> which could contain 100,000 files. And can be deleted. LCU stands
> for Last Cumulative Update (it's for an Undo function on Windows Update).

How does one locate LCU and delete it?

> Firefox has had runaway memory consumption. But when Firefox exits,
> that RAM isn't leaked and it is returned to the operating system properly.
>
> Things that leak, could be leaking at kernel level (down where the
> drivers live). Paged or unpaged pool. The problem for a Win7 user,
> is finding versions of debug tools that work on Win7 (because Microsoft
> removes stuff like that).--
Jeff Barnett

On 2/24/2024 12:52 PM, Jeff Barnett wrote:
> On 2/24/2024 8:22 AM, Paul wrote:
>>
>> As another example, before a backup you can remove the contents of LCU,
>> which could contain 100,000 files. And can be deleted. LCU stands
>> for Last Cumulative Update (it's for an Undo function on Windows Update).
>
> How does one locate LCU and delete it?

My disk here, is a dual boot W10/W11, and the directory shown is

C:\Windows\servicing\LCU # The Win10 is labeled H: at the moment

and the two "items" could be deleted (which would mean the last Patch Tuesday
could not be undone, if I were to do that now). If C: has files numbered up
to six hundred thousand , those two items contain 1/3rd of all files on the C: drive,
right now. I could just pop those two folders into the trash.

[Picture]

https://i.postimg.cc/htg0rFN9/deleting-LCU-contents.gif

It's not really the space that matters (2GB total), it's just a lot of cruft when
doing brute force searches with Agent Ransack (200,000 files). I bet most of
the audience could think of better ways to do that :-) On garbage day here,
I use a garbage bag, and do not put 200,000 items in the drive :-) Putting them
in a ZIP or a cab would be more tidy.

Another area of concern, would be the sheer quantity of bug-splat associated
with MSEdge. What a mess! Why don't they just keep a virtual copy of that
shite, on a Microsoft server. There's backup copies of backup copies of
backup copies. I don't think Internet Explorer was quite like that!

Paul

On 2/24/2024 3:44 PM, Paul wrote:
> On 2/24/2024 12:52 PM, Jeff Barnett wrote:
>> On 2/24/2024 8:22 AM, Paul wrote:
>>>
>>> As another example, before a backup you can remove the contents of LCU,
>>> which could contain 100,000 files. And can be deleted. LCU stands
>>> for Last Cumulative Update (it's for an Undo function on Windows Update).
>>
>> How does one locate LCU and delete it?
>
> My disk here, is a dual boot W10/W11, and the directory shown is
>
> C:\Windows\servicing\LCU # The Win10 is labeled H: at the moment
>
> and the two "items" could be deleted (which would mean the last Patch Tuesday
> could not be undone, if I were to do that now). If C: has files numbered up
> to six hundred thousand , those two items contain 1/3rd of all files on the C: drive,
> right now. I could just pop those two folders into the trash.
>
> [Picture]
>
> https://i.postimg.cc/htg0rFN9/deleting-LCU-contents.gif
>
> It's not really the space that matters (2GB total), it's just a lot of cruft when
> doing brute force searches with Agent Ransack (200,000 files). I bet most of
> the audience could think of better ways to do that :-) On garbage day here,
> I use a garbage bag, and do not put 200,000 items in the drive :-) Putting them
> in a ZIP or a cab would be more tidy.
>
> Another area of concern, would be the sheer quantity of bug-splat associated
> with MSEdge. What a mess! Why don't they just keep a virtual copy of that
> shite, on a Microsoft server. There's backup copies of backup copies of
> backup copies. I don't think Internet Explorer was quite like that!
Thanks for the reply. I had assumed since this is posted in the
alt.windows7.general newsgroup that this applied to Windows 7. I can
find no such directory on my 7 system.
--
Jeff Barnett

gfretwell@aol.com wrote:
> After a fresh boot I am sitting around 1.55g used. I can load Firefox
> and it gets up around 1.7. As the day goes on usage goes up, just
> browsing to over 6 spiking to the top of the box (8) with no windows
> open but I don't see any processes using that much. I am guessing a
> script has it's thumb on memory it isn't using. How do I find this
> memory hog?

Is the continued increasing with or without FF loaded?
How about when FF is never loaded to begin with?
On my w7 machines FF is mem hog and not all it's processes shut down
every time.

In message <ure6v9$1ghea$1@dont-email.me> at Sat, 24 Feb 2024 20:00:00,
Paul in Houston TX <Paul@Houston.Texas> writes
>gfretwell@aol.com wrote:
>> After a fresh boot I am sitting around 1.55g used. I can load Firefox
>> and it gets up around 1.7. As the day goes on usage goes up, just
>> browsing to over 6 spiking to the top of the box (8) with no windows
>> open but I don't see any processes using that much. I am guessing a
>> script has it's thumb on memory it isn't using. How do I find this
>> memory hog?
>
>Is the continued increasing with or without FF loaded?
>How about when FF is never loaded to begin with?

Yes, I was wondering about both of those.

>On my w7 machines FF is mem hog and not all it's processes shut down
>every time.
>
How long do you wait? I've (very occasionally) seen browser processes -
can't remember which browser - take a minute, or possibly even several,
to disappear after closing the browser.
--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

Everyone learns from science. It all depends how you use the knowledge. - "Gil
Grissom" (CSI).

On 2/24/2024 7:54 PM, Jeff Barnett wrote:

> Thanks for the reply. I had assumed since this is posted in the
> alt.windows7.general newsgroup that this applied to Windows 7.
> I can find no such directory on my 7 system.

Patch Tuesday updates stopped for Windows 7 some time ago.
And I don't know if the LCU content has an age-out provision
or not (self-cleaning).

Cumulatives did not exist for Windows 7 at first, but were
later introduced there. It could be Vista entirely missed
the feature.

It's very hard now, to memorize all this stuff, as no
feature on modern Windows lasts forever. And stuff that
was there yesterday, could be gone today.

Paul

On Sat, 24 Feb 2024 08:20:56 -0500, "Newyana2"
<Newyana2@invalid.nospam> wrote:

><gfretwell@aol.com> wrote
>
> I am guessing a
>| script has it's thumb on memory it isn't using. How do I find this
>| memory hog?
>
> Process Explorer. "Working set" seems to be the closest thing
> to actual RAM usage. Alsio perhaps check FF cache. I set it to
>about 10 MB because cache is rarely used these days. In the
>56K days browsers would cache webpages for revisits. Today,
>nearly every webpage is beiing generated "new", so there are
>few 304 server responses. And with fast speeds thheres no need
>to cache small files.
>
>There are numerous other things. Unnecessary services? AV that's
>checking everything that moves? Also, IE brings down all earlier
>systems by tying into Explorer. I'm not sure about Win7. But you can
>set temp internet files to something like 10 MB and clear the cache,
>anyway.
>

Does cache get cleared on a boot? When I reboot this think it cruises
at 1.5g or so, nothing going and it might get up to 3. I have been
browsing my normal paths today and it still drops back to 1.5.
Yesterday it was cruising at 6 and spiking into the top of the box,
sometimes bumping it's head for minutes.
I am keeping the CPU and Memory box on the side of my active window
all day watching it. My guess at this point is it must have been a
different web site I visited. Maybe I will run yesterday's history
again with the memory box up and see if that ugly thing raises it's
head again.

On Sat, 24 Feb 2024 10:22:04 -0500, Paul <nospam@needed.invalid>
wrote:

>On 2/24/2024 8:20 AM, Newyana2 wrote:
>> <gfretwell@aol.com> wrote
>>
>> I am guessing a
>> | script has it's thumb on memory it isn't using. How do I find this
>> | memory hog?
>>
>> Process Explorer. "Working set" seems to be the closest thing
>> to actual RAM usage. Alsio perhaps check FF cache. I set it to
>> about 10 MB because cache is rarely used these days. In the
>> 56K days browsers would cache webpages for revisits. Today,
>> nearly every webpage is beiing generated "new", so there are
>> few 304 server responses. And with fast speeds thheres no need
>> to cache small files.
>>
>> There are numerous other things. Unnecessary services? AV that's
>> checking everything that moves? Also, IE brings down all earlier
>> systems by tying into Explorer. I'm not sure about Win7. But you can
>> set temp internet files to something like 10 MB and clear the cache,
>> anyway.
>
>Some browsers offer an option to have a file cache on disk (default)
>or in RAM. So it could be one or the other.
>
>The cache still gets used. Because I sometimes clean it out
>before a backup (as part of general removal of cruft).
>
>As another example, before a backup you can remove the contents of LCU,
>which could contain 100,000 files. And can be deleted. LCU stands
>for Last Cumulative Update (it's for an Undo function on Windows Update).
>
>Firefox has had runaway memory consumption. But when Firefox exits,
>that RAM isn't leaked and it is returned to the operating system properly.
>
>Things that leak, could be leaking at kernel level (down where the
>drivers live). Paged or unpaged pool. The problem for a Win7 user,
>is finding versions of debug tools that work on Win7 (because Microsoft
>removes stuff like that).
>
> Paul
>

This did not go away after every Firefox process was ended.I do have a
snapshot of the process and services boxes when it is freshly booted.
I am curious what changed if anything when it fails. .

On Sat, 24 Feb 2024 17:44:09 -0500, Paul <nospam@needed.invalid>
wrote:

>On 2/24/2024 12:52 PM, Jeff Barnett wrote:
>> On 2/24/2024 8:22 AM, Paul wrote:
>>>
>>> As another example, before a backup you can remove the contents of LCU,
>>> which could contain 100,000 files. And can be deleted. LCU stands
>>> for Last Cumulative Update (it's for an Undo function on Windows Update).
>>
>> How does one locate LCU and delete it?
>
>My disk here, is a dual boot W10/W11, and the directory shown is
>
> C:\Windows\servicing\LCU # The Win10 is labeled H: at the moment
>
>and the two "items" could be deleted (which would mean the last Patch Tuesday
>could not be undone, if I were to do that now). If C: has files numbered up
>to six hundred thousand , those two items contain 1/3rd of all files on the C: drive,
>right now. I could just pop those two folders into the trash.
>
> [Picture]
>
> https://i.postimg.cc/htg0rFN9/deleting-LCU-contents.gif
>
>It's not really the space that matters (2GB total), it's just a lot of cruft when
>doing brute force searches with Agent Ransack (200,000 files). I bet most of
>the audience could think of better ways to do that :-) On garbage day here,
>I use a garbage bag, and do not put 200,000 items in the drive :-) Putting them
>in a ZIP or a cab would be more tidy.
>
>Another area of concern, would be the sheer quantity of bug-splat associated
>with MSEdge. What a mess! Why don't they just keep a virtual copy of that
>shite, on a Microsoft server. There's backup copies of backup copies of
>backup copies. I don't think Internet Explorer was quite like that!
>
> Paul

This is W/7 After that update rollup you gave us, I stopped any other
updating.

On Sat, 24 Feb 2024 20:00:00 -0600, Paul in Houston TX
<Paul@Houston.Texas> wrote:

>gfretwell@aol.com wrote:
>> After a fresh boot I am sitting around 1.55g used. I can load Firefox
>> and it gets up around 1.7. As the day goes on usage goes up, just
>> browsing to over 6 spiking to the top of the box (8) with no windows
>> open but I don't see any processes using that much. I am guessing a
>> script has it's thumb on memory it isn't using. How do I find this
>> memory hog?
>
>Is the continued increasing with or without FF loaded?
>How about when FF is never loaded to begin with?
>On my w7 machines FF is mem hog and not all it's processes shut down
>every time.

I have been on Firefox all day opening and closing it. I still get
back to 1.5g when I close everything. It is something else.

On Sun, 25 Feb 2024 02:06:11 +0000, "J. P. Gilliver"
<G6JPG@255soft.uk> wrote:

>In message <ure6v9$1ghea$1@dont-email.me> at Sat, 24 Feb 2024 20:00:00,
>Paul in Houston TX <Paul@Houston.Texas> writes
>>gfretwell@aol.com wrote:
>>> After a fresh boot I am sitting around 1.55g used. I can load Firefox
>>> and it gets up around 1.7. As the day goes on usage goes up, just
>>> browsing to over 6 spiking to the top of the box (8) with no windows
>>> open but I don't see any processes using that much. I am guessing a
>>> script has it's thumb on memory it isn't using. How do I find this
>>> memory hog?
>>
>>Is the continued increasing with or without FF loaded?
>>How about when FF is never loaded to begin with?
>
>Yes, I was wondering about both of those.
>
>>On my w7 machines FF is mem hog and not all it's processes shut down
>>every time.
>>
>How long do you wait? I've (very occasionally) seen browser processes -
>can't remember which browser - take a minute, or possibly even several,
>to disappear after closing the browser.

Once this started cruising at 6g doing nothing, it just stayed that
way. I went out to the store, came home and mounted 2 grab bars when I
got home and it was still 6g when I sat down. Virtually 0 CPU so
nobody was mining bit coins behind my back or anything.

On Sat, 24 Feb 2024 01:07:57 -0500, gfretwell@aol.com wrote:
> After a fresh boot I am sitting around 1.55g used. I can load Firefox
> and it gets up around 1.7. As the day goes on usage goes up, just
> browsing to over 6 spiking to the top of the box (8) with no windows
> open but I don't see any processes using that much. I am guessing a
> script has it's thumb on memory it isn't using. How do I find this
> memory hog?

A clean installed browser won't have any memory leak, or at least, quite low
memory leak.

Most notable memory leak came from installed browser extensions. The browser
extension architechture itself is fundamentally not a memory leaker. It's
just that nowaday programmers don't care about resource usage efficiency -
much less manage them. IOTW, it's a matter of flawed software implementation
design.

That includes bundled/integrated browser extensions. And "system addons"
which are made part of the browser's core components which aren't registered
as browser extensions, such as Google's Safe Browsing, Opera's "VPN",
integrated browser vendors' online account support, etc.

In Chromium and clone, if it's a browser extension (bundled/integrated or
not), memory usage can be seen from the browser's (Chromium's) Task Manager
which is accessible by pressing SHIFT+ESC (note: some clones disables the
Task Manager).

In Firefox, its viewable using its Performance page (`about:performance`).
However, the listed memory usage doesn't include everything (e.g. one listed
as using only 50MB, where actually, it uses 300MB). Full memory usage is
viewable from Firefox's memory auditing page (`about:memory`).

On 2/24/2024 7:55 PM, Paul wrote:
> On 2/24/2024 7:54 PM, Jeff Barnett wrote:
>
>> Thanks for the reply. I had assumed since this is posted in the
>> alt.windows7.general newsgroup that this applied to Windows 7.
>> I can find no such directory on my 7 system.
>
> Patch Tuesday updates stopped for Windows 7 some time ago.
> And I don't know if the LCU content has an age-out provision
> or not (self-cleaning).
>
> Cumulatives did not exist for Windows 7 at first, but were
> later introduced there. It could be Vista entirely missed
> the feature.
>
> It's very hard now, to memorize all this stuff, as no
> feature on modern Windows lasts forever. And stuff that
> was there yesterday, could be gone today.
I'm just noting that the last sentence above should read: "And stuff
that was here today, could be gone yesterday." But I know how difficult
it is to say anything rational about Windows versions.
--
Jeff Barnett

Google for Superfetch Technology. Windows caches Data in Memory since Vista,
making people believe it uses too much memory.

You may disable it in Services applet.

<gfretwell@aol.com> schrieb im Newsbeitrag
news:5j1jti5nchnor5r1aenc78bscrbm7t28n6@4ax.com...
> After a fresh boot I am sitting around 1.55g used. I can load Firefox
> and it gets up around 1.7. As the day goes on usage goes up, just
> browsing to over 6 spiking to the top of the box (8) with no windows
> open but I don't see any processes using that much. I am guessing a
> script has it's thumb on memory it isn't using. How do I find this
> memory hog?

On 2/24/2024 11:09 PM, gfretwell@aol.com wrote:
> On Sat, 24 Feb 2024 08:20:56 -0500, "Newyana2"
> <Newyana2@invalid.nospam> wrote:
>
>> <gfretwell@aol.com> wrote
>>
>> I am guessing a
>> | script has it's thumb on memory it isn't using. How do I find this
>> | memory hog?
>>
>> Process Explorer. "Working set" seems to be the closest thing
>> to actual RAM usage. Alsio perhaps check FF cache. I set it to
>> about 10 MB because cache is rarely used these days. In the
>> 56K days browsers would cache webpages for revisits. Today,
>> nearly every webpage is beiing generated "new", so there are
>> few 304 server responses. And with fast speeds thheres no need
>> to cache small files.
>>
>> There are numerous other things. Unnecessary services? AV that's
>> checking everything that moves? Also, IE brings down all earlier
>> systems by tying into Explorer. I'm not sure about Win7. But you can
>> set temp internet files to something like 10 MB and clear the cache,
>> anyway.
>>
>
> Does cache get cleared on a boot? When I reboot this think it cruises
> at 1.5g or so, nothing going and it might get up to 3. I have been
> browsing my normal paths today and it still drops back to 1.5.
> Yesterday it was cruising at 6 and spiking into the top of the box,
> sometimes bumping it's head for minutes.
> I am keeping the CPU and Memory box on the side of my active window
> all day watching it. My guess at this point is it must have been a
> different web site I visited. Maybe I will run yesterday's history
> again with the memory box up and see if that ugly thing raises it's
> head again.
>

You're not on the latest Firefox in a sense, because Firefox for
Windows 7 is just the ESR version. The version that runs on Windows 10,
is the release stream. The Extended Support version would receive
bug fixes (for some flavor of "bug"). but the feature set is not
supposed to change. For example, if the browser happened to not support
tabs, then an ESR version would not be expected to introduce tabs
half way through. But if tabs existed when the first ESR of the
series came out, they would be willing to backport bug fixes for
the tab thing.

If the handling of runaway memory, had not been implemented as a
feature, then perhaps that version still has some of the runaway
behavior that existed previously.

They've been very careful in the past, not to admit what
causes runaway memory behavior. There is no root causes
such as "oh, the web developer advertising code is abusing memory".
They won't come out and say that.

Even today, while I'm browsing one of the bigger sites, all
of a sudden, the fans on my PC spool up, the "stop button"
flashes on and off ten times a second, and a large number of
domains are accessed. These web monkeys have not stopped
with their monkey-shines. They're still "doing shit".

Paul

On 2/24/2024 9:00 PM, Paul in Houston TX wrote:
> gfretwell@aol.com wrote:
>> After a fresh boot I am sitting around 1.55g used. I can load Firefox
>> and it gets up around 1.7. As the day goes on usage goes up, just
>> browsing  to over 6 spiking to the top of the box (8) with no windows
>> open but I don't see any processes using that much. I am guessing a
>> script has it's thumb on memory it isn't using. How do I find this
>> memory hog?
>
> Is the continued increasing with or without FF loaded?
> How about when FF is never loaded to begin with?
> On my w7 machines FF is mem hog and not all it's processes shut down every time.
>

You should try the parent PID.

You'll need a Process Explorer for Windows 7, as the
current one might not be good for it.

https://learn.microsoft.com/en-us/sysinternals/downloads/process-explorer

When it lists Firefox, the parent PID is leftmost,
the others (children) are indented. Killing the parent
should cause all the children to be harvested.

If you kill children, their named pipe to the parent is closed,
but the parent might just restart the child (comedy ensues). We
refer to this as a flavor of Whack-A-Mole. Some children are
more restart-able than others (a failure is assumed to be an
exploit by a specially crafted video). The children are containers,
to avoid exploits, and some containers have a higher risk than
others, and the high risk containers are just restarted.

Task Manager only has a column for PID and nothing for PPID.
Whereas Process Explorer lists PPID so you know who the parent is.

When Quantum came out, the shutdown sequence for Firefox was
extremely slow (20 seconds). They were having trouble with their
management of children and runstate. They've since fixed that.
But the fastest result should come from hammering the PPID
(the parent of all the children).

Paul

On 2/24/2024 4:54 PM, Jeff Barnett wrote:
> Thanks for the reply. I had assumed since this is posted in the
> alt.windows7.general newsgroup that this applied to Windows 7. I can
> find no such directory on my 7 system.

Windows 7 Ultimate SP1 x64

I found LCU at [C:\Windows\servicing\LCU] as described in an earlier
post in this thread. Not having any recent updates to Windows 7 other
than virus definitions for Microsoft's Security Essentials, I was not
surprised to see that the most recent files in LCU were about two years
old. (Virus definitions are data and not executables and thus did not
appear in LCU.)

I had over 5,000 old files in LCU. When I tried to delete some, they
were locked. I then used Unlocker.exe from
<http://www.emptyloop.com/unlocker/>. It took a long time for
Unlocker.exe to unlock and then delete all the files.

--
David E. Ross
<http://www.rosde.com/>

Paris mayor quits X platform, calling it a 'gigantic global sewer'.
Others characterize X (previously known as Twitter) as the place
where truth goes to die.

On 2/23/2024 10:07 PM, gfretwell@aol.com wrote:
> After a fresh boot I am sitting around 1.55g used. I can load Firefox
> and it gets up around 1.7. As the day goes on usage goes up, just
> browsing to over 6 spiking to the top of the box (8) with no windows
> open but I don't see any processes using that much. I am guessing a
> script has it's thumb on memory it isn't using. How do I find this
> memory hog?
>

In Firefox's address bar, enter <about:memory>. There is an option to
reduce memory usage.

--
David E. Ross
<http://www.rossde.com/>

Demonstrators worldwide are demanding that Israel stop
fighting in Gaza. Why does it seem that no one is demanding
that Hamas stop fighting? And where are the demonstrations
against Russia fighting in the Ukraine.

On Sun, 25 Feb 2024 05:56:34 -0500, Paul <nospam@needed.invalid>
wrote:

>On 2/24/2024 11:09 PM, gfretwell@aol.com wrote:
>> On Sat, 24 Feb 2024 08:20:56 -0500, "Newyana2"
>> <Newyana2@invalid.nospam> wrote:
>>
>>> <gfretwell@aol.com> wrote
>>>
>>> I am guessing a
>>> | script has it's thumb on memory it isn't using. How do I find this
>>> | memory hog?
>>>
>>> Process Explorer. "Working set" seems to be the closest thing
>>> to actual RAM usage. Alsio perhaps check FF cache. I set it to
>>> about 10 MB because cache is rarely used these days. In the
>>> 56K days browsers would cache webpages for revisits. Today,
>>> nearly every webpage is beiing generated "new", so there are
>>> few 304 server responses. And with fast speeds thheres no need
>>> to cache small files.
>>>
>>> There are numerous other things. Unnecessary services? AV that's
>>> checking everything that moves? Also, IE brings down all earlier
>>> systems by tying into Explorer. I'm not sure about Win7. But you can
>>> set temp internet files to something like 10 MB and clear the cache,
>>> anyway.
>>>
>>
>> Does cache get cleared on a boot? When I reboot this think it cruises
>> at 1.5g or so, nothing going and it might get up to 3. I have been
>> browsing my normal paths today and it still drops back to 1.5.
>> Yesterday it was cruising at 6 and spiking into the top of the box,
>> sometimes bumping it's head for minutes.
>> I am keeping the CPU and Memory box on the side of my active window
>> all day watching it. My guess at this point is it must have been a
>> different web site I visited. Maybe I will run yesterday's history
>> again with the memory box up and see if that ugly thing raises it's
>> head again.
>>
>
>You're not on the latest Firefox in a sense, because Firefox for
>Windows 7 is just the ESR version. The version that runs on Windows 10,
>is the release stream. The Extended Support version would receive
>bug fixes (for some flavor of "bug"). but the feature set is not
>supposed to change. For example, if the browser happened to not support
>tabs, then an ESR version would not be expected to introduce tabs
>half way through. But if tabs existed when the first ESR of the
>series came out, they would be willing to backport bug fixes for
>the tab thing.
>
>If the handling of runaway memory, had not been implemented as a
>feature, then perhaps that version still has some of the runaway
>behavior that existed previously.
>
>They've been very careful in the past, not to admit what
>causes runaway memory behavior. There is no root causes
>such as "oh, the web developer advertising code is abusing memory".
>They won't come out and say that.
>
>Even today, while I'm browsing one of the bigger sites, all
>of a sudden, the fans on my PC spool up, the "stop button"
>flashes on and off ten times a second, and a large number of
>domains are accessed. These web monkeys have not stopped
>with their monkey-shines. They're still "doing shit".
>
> Paul

This seems to be a death by 1000 cuts. I have been watching the memory
tab and I haven't seen any particular site or activity spiking the
memory but my cruising total with everything stopped and the processes
page looking about the same has crept up to 2.5g. I think it may just
be a lot of different things leaving a fist full of bread crumbs as it
leaves. I guess the answer is to just boot the machine more often. It
will go back to 1.5G after a warm boot.
At least I know why this thing starts crawling in the mud after a
while. If it is paging, just sitting at the desk top you are running
at the speed of the hard drive and paging overhead, not RAM speed.

Thanks to all who responded. I always learn things here.

On 2/25/2024 3:58 PM, David E. Ross wrote:
> On 2/23/2024 10:07 PM, gfretwell@aol.com wrote:
>> After a fresh boot I am sitting around 1.55g used. I can load Firefox
>> and it gets up around 1.7. As the day goes on usage goes up, just
>> browsing to over 6 spiking to the top of the box (8) with no windows
>> open but I don't see any processes using that much. I am guessing a
>> script has it's thumb on memory it isn't using. How do I find this
>> memory hog?
>>
>
> In Firefox's address bar, enter <about:memory>. There is an option to
> reduce memory usage.
>

But typically, any time there is abnormal (OS measured) memory consumption,
what the browser sees internally doesn't seem to correlate well with
symptoms. These problems hardly ever seem to be conventional.

I got a lucky observation one day.

Running Firefox.
It was iconified on the task bar.
Notice in Task Manager, 3GB of memory is missing.
Open the Firefox window. The display in that window
is stale and out of date. The composited frames
are not making it through.

It would seem the compositing had stopped reaching Explorer (desktop owner).
And the 3GB could be the amount of memory used to hold all the
composited frames.

The memory used/wasted, may not be booked to the responsible party.

And the thing is, these problems are seen on both Windows and Linux,
which makes it less likely to be just a single OS problem. I have had
at least one Linux desktop session killed by OOM, because Firefox
was running and it eventually ate all free RAM. Then OOM was invoked,
and it kills the desktop management, rather than killing a "responsible"
application.

But as long as no one admits to what the root cause is
(the idea that compositing is a "good idea"), we'll never know
what is going on there for sure. Lots of other things composite,
without becoming wobbly.

Paul

In message <urg8n8$226gf$1@dont-email.me> at Sun, 25 Feb 2024 12:42:14,
David E. Ross <nobody@nowhere.invalid> writes
>On 2/24/2024 4:54 PM, Jeff Barnett wrote:
>> Thanks for the reply. I had assumed since this is posted in the
>> alt.windows7.general newsgroup that this applied to Windows 7. I can
>> find no such directory on my 7 system.
>
>Windows 7 Ultimate SP1 x64
>
>I found LCU at [C:\Windows\servicing\LCU] as described in an earlier
>post in this thread. Not having any recent updates to Windows 7 other
>than virus definitions for Microsoft's Security Essentials, I was not
>surprised to see that the most recent files in LCU were about two years
>old. (Virus definitions are data and not executables and thus did not
>appear in LCU.)

Windows 7 Home (I think) SP1 x86

I don't have LCU, but I do have servicing - it's only 27.8 MB in total
though - 7 subdirectories:
servicing itself TrustedInstaller.exe & 3 DLLs, 2010 & 2009
Editions 7 .xml files, 2009-6-10 & 2010-11-20, total 0.309 MB
en-US 2 .mui files, 2009-7-14, 12 and 6 KB
GC32 tzupd.exe, 2010-11-20, 40 KB
Packages 1,175 files, 2009-7-13 to 2019-8-29, total 24.7 MB
Sessions 34 .xml files, 2023-1-25 to 2024-2-6, total 2.41 MB
SQM 21 .sqm files, all 2024-2-6, all < 1 KB
Version just one directory 6.1.7601.24383 2024.2.6 containing one file
x86_installed 9 bytes ("cookie " CRLF)

The Editions folder is dated 2023-1-25 which could be when W7 was
installed on this machine (I bought it 2023-1-27); not sure what the
significance of 2024-2-6 is - possibly last time I booted (W7 is very
reliable on this machine!).
>
>I had over 5,000 old files in LCU. When I tried to delete some, they
>were locked. I then used Unlocker.exe from
><http://www.emptyloop.com/unlocker/>. It took a long time for
>Unlocker.exe to unlock and then delete all the files.
>
So I have 1,245 files, 27.8 MB, so probably not worried about them.
--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

a joke doesn't become a dad joke until it's full groan