Looking at the timestamps, I have BOX3-eth1 and BOX3-eth2 backwards.
BOX3 is doing something wrong with the return traffic, and my guess is
that its policy routing rule says to send ALL HTTP-related traffic to
BOX1. If so, the rule should be to send all traffic with a DESTINATION
port of 80 to BOX1, but route SOURCE 80 normally.

Hope that helps,
Jack

On Mon, 28 Jan 2002, Jack Coates wrote:

> Well, here's what I've got so far -- I didn't get any sleep last night
> and need to go fix that, but here's a few questions and assumptions:
>
> SYN 192.168.10.3:2727 -> eth1[BOX3]eth2 -> eth1[BOX1]ppp0
> NAT:62.234.0.234.61706 -> www.monkeynoodle.org:80
>
> packet goes into BOX3
> 06:34:16.517303 192.168.10.3.2727 > 66.1.155.123.80: S
> 1254467949:1254467949(0) win 16384 <mss 1460,nop,nop,sackOK> (DF)
> packet comes out of BOX3
> 06:34:16.517089 192.168.10.3.2727 > 66.1.155.123.80: S
> 1254467949:1254467949(0) win 16384 <mss 1460,nop,nop,sackOK> (DF)
> packet goes into BOX1 and gets NAT'd
> ASSUMPTION -- BOX1's clock is 15 seconds fast.
> packet comes out of BOX1
> 06:34:31.223667 62.234.0.234.61706 > 66.1.155.123.80: S
> 1254467949:1254467949(0) win 16384 <mss 1460,nop,nop,sackOK> (DF)
>
> 2/10ths of a second later...
> 192.168.10.3:2727 <- eth1[BOX3]eth2 <- eth1[BOX1]ppp0
> NAT:62.234.0.234.61706 <- www.monkeynoodle.org:80 ACK
>
> packet goes into BOX1 and gets NAT'd
> 06:34:31.443667 66.1.155.123.80 > 62.234.0.234.61706: S
> 3199824407:3199824407(0) ack 1254467950 win 5840 <mss
> 1412,nop,nop,sackOK> (DF)
> the BOX3-eth2 trace never shows packets coming back from the Internet,
> only leaving.
> ASSUMPTION: packet goes into BOX3
> packet comes out of BOX3
> 06:34:16.747496 66.1.155.123.80 > 192.168.10.3.2727: S
> 3199824407:3199824407(0) ack 1254467950 win 5840 <mss
> 1412,nop,nop,sackOK> (DF)
>
> I'll finish up tomorrow night, but BOX3 ETH2 is a place to start
> looking.
> Jack
>
>
> On Mon, 28 Jan 2002, Reginald R. Richardson wrote:
>
> > Ok Jack, talk to me know, have some info for you...i think we going to get it talk 
>now, i think i see the problem, but
> > the solution, i need you helping minds again...
> >
> > Attached you'll find tcpdump files of what's happening with these Routers overhere 
>in Europe..
> >
> > My understanding of the DUMP, is not up to par, but according to me this is what i 
>see and assumed, but as always, u can
> > correct if i'm wrong..
> >
> > Workstation 192.168.10.3 is sending his HTTP (80) traffic to his default router 
>Box3 (eth1) 192.168.10.254, and i can
> > clearly see him forward it according the the CABLE rule (fwmark2) to Box1 (eth), 
>so no problem there, after that short
> > journey, i see Box1 (eth1) forwards it to the Internet via ppp0, so everybody 
>happy there.......
> >
> > No the Internet "www.monkeynoodle.org" kindly accepts this request, and for some 
>reason or the other, decides to answer
> > to this poor request coming from europe......as i check again, i can see PPP0 
>telling www.monkeynoodle.org, yes, yes..i
> > sent u a request...so gimme my  reply, and he kindly answers that reply, and 
>forwards it to his next door neighbour
> > (box1 eth1), no he feels good, that he gets his reply back, and being a good guy, 
>he sends it back down the chain to
> > BOX3 eth2,
> > No box2 see this Port 80 packet coming in LOUD and clear...and kindly answers it 
>with joy, to forward it back to the
> > poor Workstation, that's waiting in vain for a reply, but eth2  has to send it via 
>his neighbour, which is BOX3 eth1,
> > which i can clearly see him doing.....
> >
> > But wait just one sec there..(Houston, i think we have a problem), yep....eth1 is 
>either refusing to answer, or he's
> > just not seeing this Port80 packet coming to him from eth1  ...TIMEOUT...RAIN 
>CHECK.....
> >
> > Now were here wondering WHAT the hell went wrong, is that, eth1 is angry with his 
>neighbour eth2 and refuse to answer,
> > or is it that he don't know the way back to send the packet back to the poor 
>workstation (192.168.10.3).
> >
> > Now, help us (me, myself and I) out there, what is missing here...well i think you 
>read my entire ip routes and ip
> > tables etc, so u have enough info to see whaz wrong, if any more info is  needed 
>please let me know and i'll send it
> > live and direct to you...
> >
> > attached u'll find tcpdumps, and somekind of ASCII netdiagram  of HomeNet in 
>Europe....struggling to offer Mommy, Daddy
> > and kids a descent internet connection..
> >
> > BTW:i was looking at leaf for the ipcheck, but ain't find it...do u have a link 
>for me...
> > thnks for the help so far..
> >
> > I think we going to get it work now.....but this is PHASE I, Phase II to follow, 
>that is PORT FORWARDING, had some
> > problems with it, but will check it out again, after we have this running like a 
>TRAIN
> >
> > Once again, thanks for your help and your ENERGY.....
> > I think i'll get this one working, i'm seeing the LIGHT, better than when i was 
>trying it with 1BOX, and two, external
> > interfaces...
> >
> > I HAVE A DREAM/HOPE, that it gonna work..
> >
> > cheers
> > Reggie
> >
> >
> > On Sat, 26 Jan 2002 15:35:55 -0800 (PST), Jack Coates wrote:
> > >On Sat, 26 Jan 2002, Reginald R. Richardson wrote:
> > >
> > >>�Jack../Charles
> > >>
> > >>�we starting to see some light, but i guess that the lack of some
> > >>Linux Firewall
> > >>�knowledge holding us back over here...
> > >>�but here's what..
> > >>
> > >>
> > >>�On my BOX3 Non NAT/Firewall Box
> > >>�if i add a default route on this box, via the CABLE Router (Box1),
> > >>then all
> > >>�HTTP traffic goes out to the internet without a problem, and also,
> > >>all the
> > >>�other traffic that has to go to the internet via Box2, goes to
> > >>Box2, so here i
> > >>�can see that Box3, is sending the traffic to the correct InterNet
> > >>Router, so in
> > >>�other words, he's a very nice Traffic Police, he's routing as
> > >>COMMANDED too..
> > >>
> > >>�For some reason, i can't figure out, why the return traffic is not
> > >>going back
> > >>�to the workstation without any problem..
> > >>
> > >
> > >To figure this out you need to use tcpdump; it's probably getting
> > >lost
> > >between box1 or 2 and box3.
> > >
> > >>�but what i found strange, is that from the moment i say the the
> > >>default gateway
> > >>�is box 1 eg.
> > >>
> > >>�"ip route add 0/0 via 192.168.1.6" (box1), then i have no problem
> > >>internet
> > >>�traffic proceeds, but from the moment i removed this route, no
> > >>more internet...
> > >>
> > >>�to the little knowledge i have, i don't believe that BOX3 should
> > >>have an
> > >>�default route, because i assume that the LOOKUP table is supposed
> > >>to tell him
> > >>�where to send the data for the specific Traffice Type. (correct me
> > >>if i'm
> > >>�wrong)
> > >>
> > >
> > >Maybe... a default route could be helpful if you get everything else
> > >configured right.
> > >
> > >>�On Box1 and Box2, is the normal settings that came by
> > >>default..with Dachsten
> > >>�onliest changes i have in those boxes is a static route back to the
> > >>�192.168.10.0 network, and i commented out the ipchains commands
> > >>that block
> > >>�traffic to the 10.0.0.0 network on Box2 (see below)
> > >>
> > >>�Box1 (Cable)
> > >>�#ip route
> > >>�62.234.0.1 dev ppp0 �proto kernel �scope link �src 62.234.0.234
> > >>�192.168.1.4/30 dev eth1 �proto kernel �scope link �src 192.168.1.6
> > >>�192.168.10.0/24 via 192.168.1.5 dev eth1
> > >>�default via 62.234.0.1 dev ppp0
> > >>
> > >>�#ip addr sh
> > >>�7: eth0: <BROADCAST,MULTICAST,UP>�mtu 1500 qdisc pfifo_fast qlen
> > >>100
> > >>���link/ether 00:10:4b:bb:c8:25 brd ff:ff:ff:ff:ff:ff
> > >>�8: eth1: <BROADCAST,MULTICAST,PROMISC,UP>�mtu 1500 qdisc
> > >>pfifo_fast qlen 100
> > >>���link/ether 00:c0:f0:12:f1:c8 brd ff:ff:ff:ff:ff:ff
> > >>���inet 192.168.1.6/30 brd 192.168.1.7 scope global eth1
> > >>
> > >>�Box2 (Adsl)
> > >>�#ip route
> > >>�192.168.1.0/30 dev eth1 �proto kernel �scope link �src 192.168.1.2
> > >>�10.0.0.0/24 dev eth0 �proto kernel �scope link �src 10.0.0.100
> > >>�192.168.10.0/24 via 192.168.1.1 dev eth1
> > >>�default via 10.0.0.138 dev eth0
> > >>
> > >>�#ip addr sh
> > >>�7: eth0: <BROADCAST,MULTICAST,UP>�mtu 1500 qdisc pfifo_fast qlen
> > >>100
> > >>���link/ether 08:00:00:22:20:34 brd ff:ff:ff:ff:ff:ff
> > >>���inet 10.0.0.100/24 brd 10.0.0.255 scope global eth0
> > >>�8: eth1: <BROADCAST,MULTICAST,PROMISC,UP>�mtu 1500 qdisc
> > >>pfifo_fast qlen 100
> > >>���link/ether 00:40:05:27:cb:9a brd ff:ff:ff:ff:ff:ff
> > >>
> > >>�This is a little tricky one, cause my ADSL provider Network
> > >>requires us to
> > >>�create a VPN connection between my router and the ADSL MODEM, so
> > >>therefore the
> > >>�default route is the ADSL Modem 10.0.0.138 (before u asked, i
> > >>commented out the
> > >>�IPCHAINS rules in this router that block the RFC ip's of 10.0.0.0)
> > >>
> > >>�>From this router i can ping the internet without any problem, so
> > >>therefore i
> > >>�have internet connectivity.
> > >>
> > >>�Here is what i have on Box3
> > >>�#ip addr sh
> > >>�7: eth0: <BROADCAST,MULTICAST,UP>�mtu 1500 qdisc pfifo_fast qlen
> > >>100
> > >>���link/ether 00:10:4b:bb:c8:25 brd ff:ff:ff:ff:ff:ff
> > >>�8: eth1: <BROADCAST,MULTICAST,PROMISC,UP>�mtu 1500 qdisc
> > >>pfifo_fast qlen 100
> > >>���link/ether 00:c0:f0:12:f1:c8 brd ff:ff:ff:ff:ff:ff
> > >>���inet 192.168.1.6/30 brd 192.168.1.7 scope global eth1
> > >>
> > >>
> > >>�# ip ru ls
> > >>�0: � � �from all lookup local
> > >>�32764: �from all fwmark � � � �1 lookup adsl
> > >>�32765: �from all fwmark � � � �2 lookup cable
> > >>�32766: �from all lookup main
> > >>�32767: �from all lookup default
> > >>
> > >>
> > >>�# ipchains
> > >>�Chain input (policy ACCEPT: 100740 packets, 8739050 bytes):
> > >>�prot opt � �tosa tosx �ifname � mark �outsize source destination
> > >>��ports
> > >>�tcp �------ 0xFF 0x00 �* � � 0x2 � �192.168.10.0/24 �0.0.0.0/0 � �*
> > >>�->� �80
> > >>�udp �------ 0xFF 0x00 �* � � 0x2 � �192.168.10.0/24 �0.0.0.0/0 � �*
> > >>�->� �80
> > >>�udp �------ 0xFF 0x00 �* � � 0x2 � �192.168.10.0/24 �0.0.0.0/0 � �*
> > >>�->� �443
> > >>�tcp �------ 0xFF 0x00 �* � � 0x2 � �192.168.10.0/24 �0.0.0.0/0 � �*
> > >>�->� �443
> > >>�tcp �------ 0xFF 0x00 �* � � 0x2 � �192.168.10.0/24 �0.0.0.0/0 � �*
> > >>�->� �110
> > >>�tcp �------ 0xFF 0x00 �* � � 0x2 � �192.168.10.0/24 �0.0.0.0/0 � �*
> > >>�->� �25
> > >>�tcp �------ 0xFF 0x00 �* � � 0x1 � �192.168.10.0/24 �0.0.0.0/0 � �*
> > >>�->� �1214
> > >>�Chain forward (policy ACCEPT: 75921 packets, 6589166 bytes):
> > >>�Chain output (policy ACCEPT: 95403 packets, 8331173 bytes):
> > >>
> > >>�# ip ro ls table cable
> > >>�default via 192.168.1.6 dev eth2
> > >>
> > >>�# ip rou ls table adsl
> > >>�default via 192.168.1.2 dev eth0
> > >>
> > >>�# ip route
> > >>�192.168.1.0/30 dev eth0 �proto kernel �scope link �src 192.168.1.1
> > >>�192.168.1.4/30 dev eth2 �proto kernel �scope link �src 192.168.1.5
> > >>�192.168.10.0/24 dev eth1 �proto kernel �scope link �src
> > >>192.168.10.254
> > >>
> > >
> > >Looks alright from a cursory glance, tcpdump is the only way to tell
> > >if
> > >it's really working like you expect.
> > >
> > >>
> > >>�Jack,
> > >>�What did u mean with this comment, don't under what u mean with
> > >>"tc"
> > >>�"Make sure you have proper tc rules for _both_ directions"
> > >>
> > >
> > >Sorry, I meant fwmark rules; tc is a tool from the same iproute2
> > >suite
> > >used for QoS.
> > >
> > >>�Do hope i have provided enough information, so that i can get
> > >>these babies talk
> > >>�to me, and do what they should do.
> > >>
> > >>�Can some one give me a tip, on what i can do to tell BOX3 that if
> > >>he routes
> > >>�HTTP traffic to BOX1, and there is no reply, then he should send
> > >>it to Box2
> > >>
> > >
> > >First concentrate on getting fwmark to work, then worry about
> > >failover
> > >:-) To do this you'll just find the ipcheck script from the LEAF site
> > >and modify it to your needs.
> > >
> > >>�thnks alot
> > >>
> > >>�On Sat, 26 Jan 2002 08:26:44 -0800 (PST), Jack Coates wrote:
> > >>�>Been there done that :-) Make sure you have proper tc rules for
> > >>�>_both_
> > >>�>directions, and try tcpdump on all three boxes. Not sure if you
> > >>�>already
> > >>�>knew this, but tcpdump has a ton of command line options to make
> > >>it
> > >>�>just
> > >>�>show the packets you're looking for. Also double-check your NAT
> > >>and
> > >>�>the
> > >>�>routing on box 1 and 2. I suspect something like this is
> > >>happening to
> > >>�>you:
> > >>�>
> > >>�>z.z.z.z:1024 SYN ->�box3 ->�box1(NATSRC=x.x.x.x:4001) ->
> > >>a.a.a.a:80
> > >>�>
> > >>�>z.z.z.z:1024 � � � �box3 <ACK loops back to>�box1 � � <-
> > >>a.a.a.a:80
> > >>�>
> > >>�>So on each box get two consoles (one for eth0 and one for eth1),
> > >>�>then do
> > >>�>a:
> > >>�>tcpdump -i eth[0|1] -n port 80 and host 66.1.155.123
> > >>�>
> > >>�>and then go to your client workstation and browse to
> > >>�>www.monkeynoodle.org. The tcpdump output should make it very clear
> > >>�>what
> > >>�>happened.
> > >>�>
> > >>�>Good luck!
> > >>�>Jack
> > >>�>
> > >>�>On Sat, 26 Jan 2002, Reginald R. Richardson wrote:
> > >>�>
> > >>�>>�Me again..
> > >>�>>
> > >>�>>�We getting there, with this 3 router box...
> > >>�>>
> > >>�>>�Question:
> > >>�>>�I reach so far as having Router3 sending the HTTP traffic to the
> > >>�>>correct
> > >>�>>�router, the SMTP traffic to the correct box also, as i use my
> > >>�>>TCPDUMP on my BOX
> > >>�>>�connecected to the Internet, i can see the HTTP traffic being
> > >>�>>transmitted to
> > >>�>>�the internet, but my problem is it's not being return to the
> > >>�>>requesting
> > >>�>>�workstation.
> > >>�>>
> > >>�>>�this is what my HTTP lookup table looks like
> > >>�>>�ip rout ls table http
> > >>�>>�default dev eth2 �scope link
> > >>�>>
> > >>�>>�I must say, that if i clear this table, and let BOX3, with a
> > >>�>>DEFAULT GW to the
> > >>�>>�internet via BOX1 or BOX2, then the Workstation can connect to
> > >>the
> > >>�>>net without
> > >>�>>�any problems.
> > >>�>>
> > >>�>>�I don't have the slightest idea now where i should look
> > >>�>>
> > >>�>>�thnks
> > >>�>>
> > >>�>>�On Wed, 23 Jan 2002 14:14:37 -0600, Charles Steinkuehler wrote:
> > >>�>>�>Everything seems to be moving like a charm, not getting the IP
> > >>�>>ROUTE
> > >>�>>�>per TCP
> > >>�>>�>Port talking to healthy, but still working on it..
> > >>�>>�>
> > >>�>>�>question.
> > >>�>>�>U mentioned why not use "equal-weight routing", i checked at
> > >>�>>googles
> > >>�>>�>to get
> > >>�>>�>more info about this, it seems a nice way to go...but can u
> > >>guide
> > >>�>>me
> > >>�>>�>to a
> > >>�>>�>weblink where i can find more info on how to implement this on
> > >>my
> > >>�>>�>Box3,
> > >>�>>�>
> > >>�>>�>CS>�Start with the Advanced Routing HOWTO, from linuxdoc.org or
> > >>�>>�>similar...if
> > >>�>>�>you get your port-based routing tables setup, you'll be over
> > >>most
> > >>�>>of
> > >>�>>�>the
> > >>�>>�>hurdles...
> > >>�>>�>
> > >>�>>�>CS>� Keep us all posted on your progress...if you get this
> > >>�>>working,
> > >>�>>�>it's the
> > >>�>>�>first step to doing the same thing cleanly with a single box.
> > >>�>>�>
> > >>�>>�>Charles Steinkuehler
> > >>�>>�>http://lrp.steinkuehler.net
> > >>�>>�>http://c0wz.steinkuehler.net (lrp.c0wz.com mirror)
> > >>�>>�>
> > >>�>>�>
> > >>�>>
> > >>�>>
> > >>�>>
> > >>�>>
> > >>�>>�-------------------------------------------------------------
> > >>�>>�Reginald R. Richardson
> > >>�>>�[EMAIL PROTECTED] on 1/26/2002
> > >>�>>
> > >>�>>
> > >>�>>
> > >>�>>�_______________________________________________
> > >>�>>�Leaf-user mailing list
> > >>�>>�[EMAIL PROTECTED]
> > >>�>>�https://lists.sourceforge.net/lists/listinfo/leaf-user
> > >>�>>
> > >>�>
> > >>
> > >>
> > >>
> > >>
> > >>�-------------------------------------------------------------
> > >>�Reginald R. Richardson
> > >>�[EMAIL PROTECTED] on 1/26/2002
> > >>
> > >>
> > >
> >
> >
> > �
> > �
> > -------------------------------------------------------------
> > Reginald R. Richardson
> > [EMAIL PROTECTED] on 1/28/2002
> >
>
>

-- 
Jack Coates
Monkeynoodle: A Scientific Venture...


_______________________________________________
Leaf-user mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/leaf-user

Reply via email to