Re: trace enque only ?

[Anton Riedl <Anton.Riedl@ei.tum.de>]

I had the same problem with enque tracing. I think I found a way how you
can do it from the tcl script (I haven't used it too much, though. But it
seems to work correctly).

To trace enqueueing at the link between nodes n0 and n1, do:

set ef [open enqueue.tr w]
set lnk [$ns link $n0 $n1]
set etr [$ns create-trace Enque $ef $n0 $n1]
$lkn add-to-head $etr

First line opens a file.
Second line finds the link object between nodes n0 and n1.
Third line creates an Enque trace object between n0 and n1 with output
file ef.
And the fourth line adds this Enque object to the head of the link
object, so everything that arrives there is recorded.

Hope it works for you.

- Anton

Frank Chu wrote:

> Thanks.  But unless I change trace.cc to output all
> those what I want, I will have to do post processing
> which consumes a lot of time and diskspace (because
> all the trace has to be output).
>
> Changing trace.cc's output format will make some of my
> previous postprocessing programs unable to get what I
> want :(
>
> Anyway, it seems this is already the best we can do.
> Thanks.
>
> --- Lloyd Wood <L.Wood@surrey.ac.uk> wrote:
> > > The question is how to trace the Enque packets
> > > only, but still didn't find a way work.
> > >
> > >  set traceobj [$ ns create-trace Enque $f $node1
> > > $node2]
> > >  $traceobj flush
> > >  $ns flush-trace
> > >
> > > It just didn't work.  What I am trying to do is to
> > let
> > > the trace file $f contains only the packets when
> > they
> > > are enqued.  However, no file was output.
> >
> > Generating a tracefile the normal way, then:
> >
> > grep '+ ' out.tr >enque.tr
> >        ^ note the space.
> >
> > works for me. Or you could hack trace.cc around. Or
> > you could
> > postprocess the tracefile with awk and only do stuff
> > if $1
> > was equal to '+'. Lots of ways to skin this
> > particular cat.
> >
> >
> > > I tried many other methods and read many times the
> > NS
> > > doc's chapter on trace and felt desperate.
> >
> > Don't worry, that much is perfectly normal. Been
> > there, done that.
> >
> > cheers,
> >
> > L.
> >
> > > I wonder
> > > if I missed anything because I am just trying to
> > do a
> > > very simple thing.  Very much appreciate if anyone
> > > can help.
> > >
> > > Frank
> >
> >
> <L.Wood@surrey.ac.uk>PGP<http://www.ee.surrey.ac.uk/Personal/L.Wood/>
> >
> >
>
> =====
>
> __________________________________________________
> Do You Yahoo!?
> Bid and sell for free at http://auctions.yahoo.com

------------------------------------------------------------------------
Anton Riedl

Lehrstuhl fuer Kommunikationsnetze
TU Muenchen, Arcisstr. 21, D-80290 Muenchen
Tel:   ++49 89 289-23506
Email: Anton.Riedl@ei.tum.de
WWW:   http://www.lkn.e-technik.tu-muenchen.de/~anton
------------------------------------------------------------------------