RH's
version of ps has historically been patched to hide the multiple kernel threads
that underlie the linixthreads approach to POSIX threads. It recognizes
the 'child' type thread and doesn't display it, unless you use the -m
parm. I would assume they would continue this, patching ps in RH9 to
ignore NPTL child threads.
If RH8
is SHOWING them, then it's likely that you ARE running NPTL on RH8, and that
version of ps is patched for linuxthreads, not NPTL.
-----Burton
-----Original Message-----In fact, it was just a 5' test.
From: Dominique Lalot [mailto:[EMAIL PROTECTED]
Sent: Wednesday, June 04, 2003 12:16 AM
To: Burton M. Strauss III
Subject: RE: [Ntop-dev] BUG on NTOP
Tne main differences I can see between the two versions of the kernel, is that the thread do not behave the same way.
RH9 update: ps ax show me one ntop process
RH8 update: ps ax show me all ntop process
As it's not more stable, I'm coming back to RH9 update to get GDB working!.
A 11:10 03/06/2003 -0500, vous avez �crit :
<snip/>
As to the glibc stuff, what worries me is this:
http://sources.redhat.com/ml/libc-alpha/2003-04/msg00103.html
Which says that the standard 2.3.2 doesn't have NPTL in it yet. But, if you
read the (RedHat) changelog:
* Wed Mar 19 2003 Jakub Jelinek <[EMAIL PROTECTED]> 2.3.2-4.80
- build as 8.0 errata
...
- don't build NPTL libraries
Which seems to indicate that the code IS in there.
I guess I just lack the confidence that they didn't do SOMETHING in there,
and that's whats causing your various and sundry problems. I've seen a
little bit myself - but I'm only a dual processor (P3) and the box doesn't
get pushed hard.
-----Burton
-----Original Message-----
From: Dominique Lalot [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, June 03, 2003 10:51 AM
To: Burton M. Strauss III
Subject: RE: [Ntop-dev] BUG on NTOP
Burton,
This is a very old message!. Thanks anyway
Dominique
