El 19/12/2010, a las 23:15, Les Mikesell <lesmikesell at gmail.com> escribió: > On 12/19/10 4:08 PM, José María Terry Jiménez wrote: >> Les Mikesell escribió: >>> On 12/19/10 2:30 PM, Jose Maria Terry Jimenez wrote: >>> >>>>>>>> This doesn't make much sense without a route. Can you try a traceroute to the >>>>>>>> fedora box address from the 192.168.236.80 box to see how/why it gets there >> >>>>>>>> Hope it helps (all addresses are 192.168. Trimmed to compact the schema): >>>>>>>> >>>>>>>> >>>>>>>> ---------- ---------- ----------- >>>>>>>> ! 1.3 !------!1.100 ! !gw 236.21! >>>>>>>> ! gw 1.1 ! ! ! 236.74!-----! 236.80 ! >>>>>>>> ---------- ! ! gw 1.1 ! ! ----------- >>>>>>>> ! ---------- ! >>>>>>>> ! ! >>>>>>>> [Router1] [Router2] >>>>>>>> >>>>>>>> Router 1 is a PFSense and its IP is 192.168.1.1 >>>>>>>> Router 2 is "something" (it is managed by other person, and i think is >>>>>>>> somekind of win server) and IP is 192.168.236.21 >>>>>>>> >>>>>>>> >>> >>> This still doesn't explain why the 192.168.236.80 box can return packets to the >>> fedora at 192.168.1.3 when you said it didn't have a route going through >>> 192.168.236.74. Can you check what routes you do have on 192.168.236.80 and >>> traceroute from there to 192.168.1.3? >>> >>> >> Apologies by confusing you. I forgot that "the other" CentOS had 2 NICs, this is >> the machine where i began these tests. It's in a remote site and now when >> listing the routes remembered that. >> >> It's conected to the 1. network with a second NIC and IP: 192.168.1.102. Replies >> must be return by that iface, really? > > Yes, with rare exceptions routing always happens with each hop making the > decision to use the interface that has the best route towards the destination, > and that would have a route automatically added for anything within the netmask. > Thanks by your help, now i understand this a bit better, Best