Behaves a bit better under AIX as well. I did have one case where a child process did
not
shutdown. Cannot repeat it though. Still this is looking pretty good.
Bill
----- Original Message -----
From: <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Thursday, June 28, 2001 9:46 PM
Subject: RE: httpd-2.0 tagged (some analysis)
>
> Glad this worked. I have been in meetings since I sent it, so I haven't
> had time to look at this yet. I will look at it as soon as I go home
> (expect about 8 pm PST). I expect this is the correct fix, and we will
> have to fix the min_spare variable. When I commit a fix, I will have a
> full explanation.
>
> Ryan
>
>
> On Thu, 28 Jun 2001, Cliff Woolley wrote:
>
> > On Thu, 28 Jun 2001, Dale Ghent wrote:
> >
> > > > That seems to work fine on HP-UX 11.0.. I'll get back on the
> > > > results on IPF (or IA64).
> > >
> > > That did the trick on Solaris 8, as well.
> >
> > Same here. I haven't taken much time to look at yesterday's change to
> > threaded.c, but this suggested fix doesn't quite "feel" right to me. If
> > we test the number of idle threads against maxspare*maxdaemons, doesn't
> > that mean that we can have way more idle threads per process available
> > than we should if we have less than the max number of daemons running
> > currently? Also, there's a test that follows that which checks the
> > number of idle threads against minspare... something needs to get done to
> > that too, I'm guessing.
> >
> > I will have to take more time to absorb these latest changes...
> >
> > --Cliff
> >
> >
> > --------------------------------------------------------------
> > Cliff Woolley
> > [EMAIL PROTECTED]
> > Charlottesville, VA
> >
> >
> >
> >
>
>
> _____________________________________________________________________________
> Ryan Bloom [EMAIL PROTECTED]
> Covalent Technologies [EMAIL PROTECTED]
> -----------------------------------------------------------------------------
>
