On Wed, Mar 24, 2010 at 11:08:04AM -0700, Darren Reed wrote:
> On Wed, 24 Mar 2010 08:37 -0700, "Ray Van Dolson" <[email protected]>
> wrote:
> > On Wed, Mar 24, 2010 at 12:16:17AM -0700, Darren Reed wrote:
> > > On Tue, 23 Mar 2010 09:14 -0700, "Ray Van Dolson" <[email protected]>
> > > wrote:
> > > > I have a multihomed box running Solaris 10 U8 (IP filter v4.1.9).
> > > > There are two interfaces, igb0 and igb2, both on the same subnet
> > > > (10.49.0.0/16) with, obviously, different IP's.
> > > > 
> > > >     igb0: 10.49.2.110/16
> > > >     igb2: 10.49.2.111/16
> > > > 
> > > >     Default Gateway: 10.49.254.254
> > > > 
> > > > When traffic destined for 10.49.2.111 enters igb2, by default replies
> > > > go back out igb0.
> > > > 
> > > > I want anything with a source IP of 10.49.2.111 to go out igb2.
> > > > 
> > > > The following two rules work:
> > > > 
> > > >   (1) block out log quick on igb0 to igb2:10.49.254.254 from 10.49.2.111
> > > >   to any
> > > >   (2) pass out log quick on igb0 to igb2:10.49.254.254 from 10.49.2.111
> > > >   to any
> > > > 
> > > > But the downside is, if the destination is also on the 10.49.0.0/16,
> > > > when it arrives it appears as if it's coming from the gateway instead
> > > > of from the MAC address of igb2.
> > > > 
> > > > I tried the following:
> > > > 
> > > >   (1) block out log quick on igb0 to igb2 from 10.49.2.111 to any
> > > >   (2) pass out log quick on igb0 to igb2 from 10.49.2.111 to any
> > > > 
> > > > But, while these rules don't complain and seem to show matches in the
> > > > log, the packets never reach the destination.
> > > > 
> > > > Any suggestions?  Do I _have_ to specify a next-hop?  I just want the
> > > > system to rely on its local ARP table for delivery, especially if the
> > > > packet is destined to the local subnet...
> > > 
> > > Are you able to use snoop/tcpdump to degtermine if anything is sent
> > > out igb2 or does the packet simply disappear down a black hole?
> > > 
> > 
> > Black hole... :)  From responses I've gotten on the OpenSolaris network
> > list, it sounds like this sort of thing won't work -- it's more aimed
> > at boxes multihomed on different subnets so I can make use of a
> > gateway.
> 
> Are you at all proficient with dtrace?
> 
> Darren
> 

I'm not unfortunately.  Guessing you'd like me to trace where the
packets ultimately end up?

At this point we've worked around the issue by adjusting our
infrastructure, but I could probably set up a test box to test this
very setup again.

Ray

Reply via email to