Re: Re:What's the longest time before Windows flushes data to disk?

Hi Slava

Concerning the "absurdities that I have withdrwawn", the only thing I can
say is that I am able to admit my mistakes - unfortunately, you seem to lack
such ability completely (in fact, this is your problem, rather than mine).
Actually, the only reason why I participate in these discussions is because
I want to learn more, so that I am not afraid of mistakes. Apparently, your
reason for participation is somehow different.....


Concerning your statements in Latin, I believe this is just pathetic
show-off - you try to demonstrate your intelligence, but somehow demonstrate
lack of any (exactly like anyone who tries to demonstate anything when
he/she does not really have to). Again, this is your problem, rather than
mine - I just find it amusing.


Concerning the faults in logic..... in order to back up your claim, you have
provided description of REALTIME_PRIORITY_CLASS from SetPriorityClass()MSDN
documentation. However, it really depends on the conclusions that you make.
For example, the only conclusion that I made from this description is that
you should use this feature sparingly and for the short intervals only,
unless you are desperate to screw up the session. However, your conclusion
is somehow different - you presented it as absolutely normal behaviour.

It is understandable that, under some certain circumstances, neither "lazy
writer" nor Balance Set Manager
may run(in fact, it applies to ANY thread, regardless of its priority). For
example, if the machine is uni-processor one and someone writes something
like

_ASM CLI

while(1){}

then they are not going to run until the reboot. I believe that your
assumption about misused REALTIME_PRIORITY_CLASS is from the same field,
although with less disastrous consequences. Therefore, I want to make it
clear that everything I said applies ONLY UNDER THE NORMAL CIRCUMSTANCES!!!!
However, you are talking about ABNORMAL(!!!) things that should not occur if
all programs that run on the target machine are properly designed, and
present them as something normal - you presented REALTIME_PRIORITY_CLASS as
one of the reasons why cache may not get flushed for minutes, hours and days.
Whose logic is at fault?????



Concerning the frequency of "lazy writer" invocation, please consult Windows
Internals, page 683, last paragraph

Concerning Balance Set Manager, please consult the same source, page 466 (I
have checked my spelling, so that you can go right to these pages).


I hope to hear from you soon - I find this "confrontation" quite
entertaining


Regards

Anton Bassov
.

Relevant Pages

  • Re: M.I.T. = Overrated Academics
    ... Oooooh.....firing someone over grammar mistakes in a memo. ... simple mistakes generally don't interfere with the communication ... I see no reason why he couldn't do both. ...
    (rec.sport.football.college)
  • Re: Why do the Mormons bother with this newsgroup?
    ... >> those features, then I would be interested to hear of them. ... make it clear that God is imposing whatever it is. ... > perfect and would have been less perfect without the mistakes. ... the narrator is a poor speller for some reason important ...
    (uk.religion.christian)
  • Re: M.I.T. = Overrated Academics
    ... Oooooh.....firing someone over grammar mistakes in a memo. ... simple mistakes generally don't interfere with the communication ... I see no reason why he couldn't do both. ... What bothers me about this thread is that we are talking about ...
    (rec.sport.football.college)
  • Re: M.I.T. = Overrated Academics
    ... Oooooh.....firing someone over grammar mistakes in a memo. ... simple mistakes generally don't interfere with the communication ... I see no reason why he couldn't do both. ...
    (rec.sport.football.college)
  • Re: M.I.T. = Overrated Academics
    ... Oooooh.....firing someone over grammar mistakes in a memo. ... simple mistakes generally don't interfere with the communication ... I see no reason why he couldn't do both. ...
    (rec.sport.football.college)