[vdr] Re: VDR deletes fresh recordings

[Rainer Zocholl <UseNet-Posting-Nospam-74308-@zocki.toppoint.de>]

vdr@zastrow4u.de(Alfred Zastrow)  15.01.05 16:56

>Rainer Zocholl schrieb:

>> How could that have happened?

>What about setting up a new system 

That IS a new system! ;-)
My beautyful almost 4 year old low power, low noise Celesius VDR 
did not like the last hot summer ;-(


>or taking  a end user suitable linux distribution. 

I have no problem with debian.

>Or replacing your hardware.

>Take a look back in the message archive. Your mails stands for
>crashes, data losses and dissatisfaction.

That is your very selective impression.
I too take the time to mail if something works.



>My system is heavily used (24/7 for a whole family), 

You are recording your family in one big 24h*7 recording? ;-)
My problem was a recording running over more than 15h...
a normal length recording would have been deleted without 
any suspisious traces left. I assume that such recordings
are not done very often.


>I have three disks and _NO_ problems, no crashes ans no data losses.

...you are arware of...

As long as there are no software problems i don't
have no problems either ;-)

As long i had used 1.2x i too had no problems (and not time to
care or write about).
(When you look into the archive you will find me under the 
"(more or less)early adopthers" at least back to 2001 with a big 
gap in the last 2 years...).

Upto now i could trace back almost every
 "crashes, data losses and dissatisfaction"
to a software error.

To i made/initiated a patch for the bad thread awaresness of 1.3.17.
Without it, vdr was unusable on "other than Klaus' system".

Such "spurious errors" caused by reentrance problems 
can often be solved^W hidden by using another hardware!
That will second everyone with experience in the multitthreading/realtime
area. Those apps are very sensitive to bugs which are not
fixed 100% waterproof. "It will mostly work" it not allowed
in threaded enviroment (but is general no good programming style IMHO).


BOT:
Klaus meanwhile found (most of) the reason:
An unclean "mostly enough" handling of deletes by the software,
combinded with creative user input,
combined with an extra long recording that give the error a chance
to occure ;-(

(If the recording were not so extrem long(approx. 15h!), i assume that this
problem would be only: "Timer did not start".)



Rainer