Check Files Problem

[OREALLY]

Probably will crash the system!

"Char Jackson" wrote in message

Yeah...stops at 517 and just keeps eating memory!

Of course, memory is there to be used, so what happens if you just let
it run? So far, all I think I've seen is that you've shut it down when
it reached a point where you weren't comfortable, but I wonder what
happens if you ignore it?

 

[Char Jackson]

"Char Jackson" wrote in message



Of course, memory is there to be used, so what happens if you just let
it run? So far, all I think I've seen is that you've shut it down when
it reached a point where you weren't comfortable, but I wonder what
happens if you ignore it?

Probably will crash the system!

I was trying to get past the 'probably' and on to the 'definite'.

Watching a program allocate available memory reminds me of one of my
customers for whom deleting a file takes three steps:
1. Delete the file.
2. Navigate to the Recycle Bin to make sure it's there.
3. Empty the Recycle Bin ("to make room for the next thing".

Anyway, if a program allocates a bunch of memory that is otherwise
just sitting there unused, my reaction is, "So what?" That's what
memory is for, and if the program eventually crashes the system,
reboot and come back to let us know because crashing the system isn't
normal, but using available memory is perfectly normal.

 

[Paul]

I was trying to get past the 'probably' and on to the 'definite'.

Watching a program allocate available memory reminds me of one of my
customers for whom deleting a file takes three steps:
1. Delete the file.
2. Navigate to the Recycle Bin to make sure it's there.
3. Empty the Recycle Bin ("to make room for the next thing".

Anyway, if a program allocates a bunch of memory that is otherwise
just sitting there unused, my reaction is, "So what?" That's what
memory is for, and if the program eventually crashes the system,
reboot and come back to let us know because crashing the system isn't
normal, but using available memory is perfectly normal.

I'm puzzled by all the comments I can find on this subject, that
engineering anything to use all available resources is a good
thing. Windows is a multi-account system, with stuff which can
be scheduled to start at any time. Inadvertently, a user could
be running a CHKDSK, at the same time some other (hog) process has
been launched. Perhaps the combination of the two, results
in extremely sluggish performance, with neither process finishing
when desired (i.e. the CHKDSK you wanted done overnight, is not
complete).

If this was such a good engineering call, perhaps CHKDSK could
shut down the OS completely, evict all other processes, and run
exclusively itself. Then, there would be no argument about
the engineering call. If CHKDSK owned the entire machine, like
sitting in an MSDOS black windows, chomping on your disk drive,
you'd be in no position to argue with it, or do other work. There
would then be no questions in your mind, about memory usage,
because there is basically nothing you can do...

At least on the Unix boxes I used years ago, there was an easy
thing to use called a memory quota. You could set it as a user,
so that the processes you were forking, could not "run away" on
you. In the situation OREALLY is in, you could on the Unix box,
set the quota to half available memory, and then the other half
of memory can be used for other things. That leaves room for
all the processes which may be automatically scheduled to start,
that you don't have control over.

The quota scheme was "per process", and it would not prevent
a system from being bogged down. If you set the limit to half
of the available memory, and launched three programs that
went to the limit, there would be swapping. But what that
limiter did do for you, is give you control over the exact
kind of behavior CHKDSK is exhibiting. Namely, the ability
to take a program with "infinite" appetite, and determine
"what it could eat". If you know a single program has "bad behavior",
the quota setting for memory would take care of it.

Windows may have a similar functionality, but it isn't nearly
as easy to get to. I haven't seen the exact registry entry
yet, to do that. Maybe some policy editor has a setting for it.

Paul

 

[Char Jackson]

I'm puzzled by all the comments I can find on this subject, that
engineering anything to use all available resources is a good
thing. Windows is a multi-account system, with stuff which can
be scheduled to start at any time. Inadvertently, a user could
be running a CHKDSK, at the same time some other (hog) process has
been launched. Perhaps the combination of the two, results
in extremely sluggish performance, with neither process finishing
when desired (i.e. the CHKDSK you wanted done overnight, is not
complete).

If this was such a good engineering call, perhaps CHKDSK could
shut down the OS completely, evict all other processes, and run
exclusively itself. Then, there would be no argument about
the engineering call. If CHKDSK owned the entire machine, like
sitting in an MSDOS black windows, chomping on your disk drive,
you'd be in no position to argue with it, or do other work. There
would then be no questions in your mind, about memory usage,
because there is basically nothing you can do...

At least on the Unix boxes I used years ago, there was an easy
thing to use called a memory quota. You could set it as a user,
so that the processes you were forking, could not "run away" on
you. In the situation OREALLY is in, you could on the Unix box,
set the quota to half available memory, and then the other half
of memory can be used for other things. That leaves room for
all the processes which may be automatically scheduled to start,
that you don't have control over.

The quota scheme was "per process", and it would not prevent
a system from being bogged down. If you set the limit to half
of the available memory, and launched three programs that
went to the limit, there would be swapping. But what that
limiter did do for you, is give you control over the exact
kind of behavior CHKDSK is exhibiting. Namely, the ability
to take a program with "infinite" appetite, and determine
"what it could eat". If you know a single program has "bad behavior",
the quota setting for memory would take care of it.

Windows may have a similar functionality, but it isn't nearly
as easy to get to. I haven't seen the exact registry entry
yet, to do that. Maybe some policy editor has a setting for it.

Windows 7 seems to be pretty good about managing its memory so I
haven't found a need to get in the middle of that. I'm not seeing what
the problem is...

 

[Paul]

Windows 7 seems to be pretty good about managing its memory so I
haven't found a need to get in the middle of that. I'm not seeing what
the problem is...

I've seen reports, that all this memory it's grabbing, isn't actually
making it run any faster. So what's the point ? Other than to
annoy the users ?

If any user reports their machine as "sluggish", what does that tell
you about how effective this design is ? Chkdsk could very well be
feeling the "sluggish" as well, and running slower than if
memory hogging wasn't a design parameter.

I've seen this kind of wrong-headed thinking in the Linux world too.
They changed "grep", a very important utility for operating a Linux
or Unix box. They set it up, to store *everything* in memory. They
also neutered a command line flag that made the behavior optional
(so you're stuck with the way it works). Maybe Microsoft was taking
lessons ?

By making that change to grep, I was forced to write my own grep analog,
so I could finish grepping a 10GB file (i.e. wouldn't fit in system
memory). Grep traditionally had a small footprint, and could be used
in command pipes. As it stands now, it's pretty useless because of
the limitations caused by "abuse of memory".

I think Microsoft at least should be showing us a performance graph,
how grabbing 7.9GB of your 8GB machine makes chkdsk run X percent faster.
Then, we could discuss... engineering. As it stands, we have to rely on
some users, to do some expensive (timewise) tests, to determine
whether this really is an advantage or not.

Just think about what a disk checking program would do. Is there
any reason to hold more than the contents of *one* file in memory ?
I don't get it. I'd like to see a statement about how the memory
is used, what their purpose is in using the memory. Is it a file
cache ? Is it a sector cache ? Does it hold speculative prefetch ?
All we got from Microsoft was "trust us" and "no, it's not a bug".

Paul

 

[Char Jackson]

I've seen reports, that all this memory it's grabbing, isn't actually
making it run any faster. So what's the point ? Other than to
annoy the users ?

If any user reports their machine as "sluggish", what does that tell
you about how effective this design is ? Chkdsk could very well be
feeling the "sluggish" as well, and running slower than if
memory hogging wasn't a design parameter.

"Reports" from users are worth what you pay for them. "Sluggish" and
"snappy" aren't technical terms, and therefore aren't very actionable.

As for people who sit there and watch memory allocation happen in real
time, I group them with the people who check and empty the Recycle Bin
after every delete and the people who muck around in Program Files and
Program Files (x86) and the Windows folder, etc. You're kind to humor
them, but my advice is to get the heck out of all of those places. In
the end, who cares how much of the available RAM a program allocates?

I think Microsoft at least should be showing us a performance graph,
how grabbing 7.9GB of your 8GB machine makes chkdsk run X percent faster.
Then, we could discuss... engineering. As it stands, we have to rely on
some users, to do some expensive (timewise) tests, to determine
whether this really is an advantage or not.

As stated above, I think user tests are mostly worthless.

This whole thread has been silly. A program that the user initiated
grabbed a lot of memory, memory that was otherwise available and
unused, so he terminated the program rather than let it run. Silly
isn't the word I really want to use, but it's Monday and I want to
start the week nicely.

In addition, I don't think Microsoft is under any obligation
whatsoever to demonstrate why using X amount of RAM is Y% better than
using some other amount.

Just think about what a disk checking program would do. Is there
any reason to hold more than the contents of *one* file in memory ?
I don't get it. I'd like to see a statement about how the memory
is used, what their purpose is in using the memory. Is it a file
cache ? Is it a sector cache ? Does it hold speculative prefetch ?
All we got from Microsoft was "trust us" and "no, it's not a bug".

If I were writing a disk checking program, it wouldn't hold the
contents of any file in memory. I don't see the point.

 

[Gene E. Bloch]

Watching a program allocate available memory reminds me of one of my
customers for whom deleting a file takes three steps:
1. Delete the file.
2. Navigate to the Recycle Bin to make sure it's there.
3. Empty the Recycle Bin ("to make room for the next thing".

Looks like a plan. I think I'll adopt it. It's so ... tidy.

 

[Char Jackson]

Looks like a plan. I think I'll adopt it. It's so ... tidy.

She's German, which sort of makes sense.

 

[Gene E. Bloch]

She's German, which sort of makes sense.

Ah, danke, das hilft mir.

 

[Ed Cryer]

She's German, which sort of makes sense.

Humour her! Tell her about this freebie minibin. She'll love it, and can
tweak it to her heart's content.
http://www.softpedia.com/reviews/windows/minibin-Review-189492.shtml

Ed

 

[Ed Cryer]

Humour her! Tell her about this freebie minibin. She'll love it, and can
tweak it to her heart's content.
http://www.softpedia.com/reviews/windows/minibin-Review-189492.shtml

Ed

Here's something to add to it;
http://freeraser.en.softonic.com/

She might have a ball configuring it to work in tandem with the Recycle
Bin; or even as a receptacle for stuff deleted from the RB.

Ed

 

[Char Jackson]

Here's something to add to it;
http://freeraser.en.softonic.com/

She might have a ball configuring it to work in tandem with the Recycle
Bin; or even as a receptacle for stuff deleted from the RB.

Thanks, Ed, but part of being German is apparently the desire to keep
extra icons off the desktop, and part of being senior (in her case,
anyway) is to keep things the same from day to day. She notices even
the most minor change and calls or emails me about it.

 

[Gene E. Bloch]

Here's something to add to it;
http://freeraser.en.softonic.com/

She might have a ball configuring it to work in tandem with the Recycle
Bin; or even as a receptacle for stuff deleted from the RB.

Ed

Geez, Ed, with all those nifty toys you're telling us about, I might
have to give up my habit of using Shift-Delete.

Actually, I do use a program similar to Freeraser to clean up a computer
when I am plan to give it to someone, so that what's on it is the
licensed Windows with nothing that can be traced to me

 

[Ed Cryer]

Geez, Ed, with all those nifty toys you're telling us about, I might
have to give up my habit of using Shift-Delete.

Actually, I do use a program similar to Freeraser to clean up a computer
when I am plan to give it to someone, so that what's on it is the
licensed Windows with nothing that can be traced to me

Some people delight in buying a new computer and then tailoring the OS
to look like the last one they had.
Others like to just tweak it to the full and produce something unique
that looks to the casual observer like nothing they've ever seen before.
And then others, like me, try and fight against both those urges and put
up with what MS in its wisdom has given us.

Ed

 

[Paul]

Some people delight in buying a new computer and then tailoring the OS
to look like the last one they had.
Others like to just tweak it to the full and produce something unique
that looks to the casual observer like nothing they've ever seen before.
And then others, like me, try and fight against both those urges and put
up with what MS in its wisdom has given us.

Ed

I like to tweak it enough, so that I'm productive.

I'm finding on my Windows 7 laptop, that my productivity is
affected a lot, by needing to stare at text too much. The laptop
usually isn't in my lap, and the screen angle is such, I don't really
have good contrast. And there's such a reliance on text for everything,
that I really have to squint at it at times. So if I had the energy for
tweaking, it would be to do something so I'd need less text.

So if there was an incentive for a more "classic" look, it would be
to end up with something that requires "less squinting" I find the
laptop screen needs to be leaned back quite a bit, to make it look good,
something there usually isn't room to do.

Paul