ipv4 connection problems

David Rhodus sdrhodus at gmail.com
Wed Mar 30 05:22:33 PST 2005


On Wed, 30 Mar 2005 02:11:46 -0700, Peter Avalos <pavalos at xxxxxxxxxxxx> wrote:
> On Sun, Mar 27, 2005 at 12:55:52PM -0800, Matthew Dillon wrote:
> >
> > :I upgraded to a kernel dated March 23rd and I have yet to reproduce the
> > :problem.  The behavior was so hard to reproduce earlier, I was going to
> > :give it a few more days to actually see if it was gone (it's hard to
> > :prove a negative).  Would it benefit you more if I upgraded to the
> > :latest kernel now, or wait a few more days to see if the bug is already
> > :fixed?  I think it'd be better to upgrade now, but wanted to see what
> > :you think, since I'm the only one who has seen this problem.
> > :
> > :--Peter
> >
> >     I would upgrade right now so we are all on the same page, just in case
> >     the clock fix was also causing the problem you report.
> >
> >     Note that crater's clock stopped this morning, and apache stopped
> >     accepting connections.  Crater had been up 50+ days so this adds
> >     some fuel to the possibility that the clock issue was causing your
> >     problem too.
> >
> >                                               -Matt
> 
> Okay, sources as of March 27th, problem still exists on an SMP kernel.  I'm
> going to give a UP kernel a shot.  I'm still trying to get a kernel core,
> but it's difficult to do since it's hard to reproduce and be at a console
> at the same time.
> 
> --Peter
> 
> 
> 

I haven't been tracking this since it started out because so many eyes
jumped right to it.  Peter, is this an example of the problem you are
seeing,  random connections to apache running on 80 don't connect, but
if you try to connect right back the, connection succeeds ?

We had a problem just like this a few months ago but it was related
somewhat to our handling of IPv6 packets ver's IPv4.  I wonder if that
could still be the case.

-- 
                                            -David
                                            Steven David Rhodus
                                            <drhodus at xxxxxxxxxxx>





More information about the Bugs mailing list