Skip to contentSkip to conversation

Problem with routing decisions, and multihop

DISCUSSION Lluís Batlle 2 participants 11 messages Jul 4, 2005 – Jul 5, 2005 1 message not in the archive

lore
Share & tools
lore
#0 lore raw

Hi!

I have many problems getting this thing to work. There's a host with two network interfaces, where there are two routers to Internet in two separated networks. The host uses multihop routing for deciding to which router send the packets... but the routing decision is wrong made. Some packets with source address of one NIC, go to other network.

I have a host with three NICs in it:
eth0 - LAN, 192.168.0.0/20
eth1 192.168.16.1 - subnetwork 192.168.16.0/28, with a router (192.168.16.2) to internet
eth2 192.168.17.1 - subnetwork 192.168.17.0/28, with another router (192.168.17.2) to internet

The routing rules are:

0:      from all lookup local 
50:     from all lookup main 
201:    from 192.168.17.0/28 iif eth2 lookup 201 
202:    from 192.168.16.0/28 iif eth1 lookup 202 
222:    from all lookup 222 
32766:  from all lookup main 
32767:  from all lookup default 

The table 'main':

192.168.17.0/28 dev eth2  proto kernel  scope link  src 192.168.17.1 
192.168.16.0/28 dev eth1  proto kernel  scope link  src 192.168.16.1 

192.168.0.0/20 dev eth0 proto kernel scope link src 192.168.1.2

The table '201':
default via 192.168.17.2 dev eth2 proto static src 192.168.17.1 prohibit default proto static metric 1

The table '202':
default via 192.168.16.2 dev eth1 proto static src 192.168.16.1 prohibit default proto static metric 1

The table '222', where there is the multihop gateway specification:

default equalize

        nexthop via 192.168.16.2  dev eth1 weight 1
        nexthop via 192.168.17.2  dev eth2 weight 1

I've added the following packet LOG lines into 'mangle' table, for knowing when the "WRONG INTERFACE" decision is being made:
Chain POSTROUTING (policy ACCEPT 329K packets, 93M bytes)
pkts bytes target prot opt in out source
destination

    2    80 LOG        all  --  any    eth1    192.168.17.1        
anywhere            LOG level warning ip-options prefix `WRONG IFACE:
'
    0     0 LOG        all  --  any    eth2    192.168.16.1        

anywhere LOG level warning ip-options prefix `WRONG IFACE:
'

(Don't look at counters; right now, for getting good internet access, I'm not using multihop)

So, often appears in the kernel log, specially with 'ftp' and 'ssh' connections (and rarely with www connections):

Jul  4 15:50:14 thecrow WRONG IFACE: IN= OUT=eth2 SRC=192.168.16.1

DST=216.165.191.52 LE
N=72 TOS=0x00 PREC=0x00 TTL=64 ID=9582 DF PROTO=TCP SPT=56528 DPT=6667 WINDOW=18824 RES=
0x00 ACK PSH URGP=0

Jul  4 16:01:29 thecrow WRONG IFACE: IN= OUT=eth1 SRC=192.168.17.1

DST=130.206.1.5 LEN=40 TOS=0x00 PREC=0x00 TTL=64 ID=66 DF PROTO=TCP SPT=33820 DPT=21 WINDOW=0 RES=0x00 RST URGP=0

Even though, when I use 'tcpdump' for catching the wrong packets (that
is: tcpdump -i eth1 host 192.168.17.1 _or_ tcpdump -i eth2 host 192.168.16.1 ) results that _A LOT MORE PACKETS_ are BADLY ROUTED, than sent to the LOG target.

My conclusion: iptables 'matching' doesn't work; also does the route decision part.

I absolutely don't know what more to do... I'm running iptables v1.2.11, and kernel 2.6.11-gentoo-r11. Exactly same happened with kernel 2.4.28-gentoo. :(

I even attach the scripts I use for doing the routing and the NAT for the LAN.

Please, help!

#1 lore raw

Hi!

I have many problems getting this thing to work. There's a host with two network interfaces, where there are two routers to Internet in two separated networks. The host uses multihop routing for deciding to which router send the packets... but the routing decision is wrong made. Some packets with source address of one NIC, go to other network.

I have a host with three NICs in it:
eth0 - LAN, 192.168.0.0/20
eth1 192.168.16.1 - subnetwork 192.168.16.0/28, with a router (192.168.16.2) to internet
eth2 192.168.17.1 - subnetwork 192.168.17.0/28, with another router (192.168.17.2) to internet

The routing rules are:

0:      from all lookup local
50:     from all lookup main
201:    from 192.168.17.0/28 iif eth2 lookup 201
202:    from 192.168.16.0/28 iif eth1 lookup 202
222:    from all lookup 222
32766:  from all lookup main
32767:  from all lookup default

The table 'main':

192.168.17.0/28 dev eth2  proto kernel  scope link  src 192.168.17.1
192.168.16.0/28 dev eth1  proto kernel  scope link  src 192.168.16.1

192.168.0.0/20 dev eth0 proto kernel scope link src 192.168.1.2

The table '201':
default via 192.168.17.2 dev eth2 proto static src 192.168.17.1 prohibit default proto static metric 1

The table '202':
default via 192.168.16.2 dev eth1 proto static src 192.168.16.1 prohibit default proto static metric 1

The table '222', where there is the multihop gateway specification:

default equalize

        nexthop via 192.168.16.2  dev eth1 weight 1
        nexthop via 192.168.17.2  dev eth2 weight 1

I've added the following packet LOG lines into 'mangle' table, for knowing when the "WRONG INTERFACE" decision is being made:
Chain POSTROUTING (policy ACCEPT 329K packets, 93M bytes)
pkts bytes target prot opt in out source
destination

    2    80 LOG        all  --  any    eth1    192.168.17.1
anywhere            LOG level warning ip-options prefix `WRONG IFACE:
'
    0     0 LOG        all  --  any    eth2    192.168.16.1

anywhere LOG level warning ip-options prefix `WRONG IFACE:
'

(Don't look at counters; right now, for getting good internet access, I'm not using multihop)

So, often appears in the kernel log, specially with 'ftp' and 'ssh' connections (and rarely with www connections):

Jul  4 15:50:14 thecrow WRONG IFACE: IN= OUT=eth2 SRC=192.168.16.1

DST=216.165.191.52 LE
N=72 TOS=0x00 PREC=0x00 TTL=64 ID=9582 DF PROTO=TCP SPT=56528 DPT=6667 WINDOW=18824 RES=
0x00 ACK PSH URGP=0

Jul  4 16:01:29 thecrow WRONG IFACE: IN= OUT=eth1 SRC=192.168.17.1

DST=130.206.1.5 LEN=40 TOS=0x00 PREC=0x00 TTL=64 ID=66 DF PROTO=TCP SPT=33820 DPT=21 WINDOW=0 RES=0x00 RST URGP=0

Even though, when I use 'tcpdump' for catching the wrong packets (that
is: tcpdump -i eth1 host 192.168.17.1 _or_ tcpdump -i eth2 host 192.168.16.1 ) results that _A LOT MORE PACKETS_ are BADLY ROUTED, than sent to the LOG target.

My conclusion: iptables 'matching' doesn't work; also does the route decision part.

I absolutely don't know what more to do... I'm running iptables v1.2.11, and kernel 2.6.11-gentoo-r11. Exactly same happened with kernel 2.4.28-gentoo. :(

I even attach the scripts I use for doing the routing and the NAT for the LAN.

Please, help!

#2 lore raw
In reply to Lluís Batlle

Hi!

I have many problems getting this thing to work. There's a host with two network interfaces, where there are two routers to Internet in two separated networks. The host uses multihop routing for deciding to which router send the packets... but the routing decision is wrong made. Some packets with source address of one NIC, go to other network.

I have a host with three NICs in it:
eth0 - LAN, 192.168.0.0/20
eth1 192.168.16.1 - subnetwork 192.168.16.0/28, with a router (192.168.16.2) to internet
eth2 192.168.17.1 - subnetwork 192.168.17.0/28, with another router (192.168.17.2) to internet

The routing rules are:

0:      from all lookup local
50:     from all lookup main
201:    from 192.168.17.0/28 iif eth2 lookup 201
202:    from 192.168.16.0/28 iif eth1 lookup 202
222:    from all lookup 222
32766:  from all lookup main
32767:  from all lookup default

The table 'main':

192.168.17.0/28 dev eth2  proto kernel  scope link  src 192.168.17.1
192.168.16.0/28 dev eth1  proto kernel  scope link  src 192.168.16.1

192.168.0.0/20 dev eth0 proto kernel scope link src 192.168.1.2

The table '201':
default via 192.168.17.2 dev eth2 proto static src 192.168.17.1 prohibit default proto static metric 1

The table '202':
default via 192.168.16.2 dev eth1 proto static src 192.168.16.1 prohibit default proto static metric 1

The table '222', where there is the multihop gateway specification:

default equalize

        nexthop via 192.168.16.2  dev eth1 weight 1
        nexthop via 192.168.17.2  dev eth2 weight 1

I've added the following packet LOG lines into 'mangle' table, for knowing when the "WRONG INTERFACE" decision is being made:
Chain POSTROUTING (policy ACCEPT 329K packets, 93M bytes)
pkts bytes target prot opt in out source
destination

    2    80 LOG        all  --  any    eth1    192.168.17.1
anywhere            LOG level warning ip-options prefix `WRONG IFACE:
'
    0     0 LOG        all  --  any    eth2    192.168.16.1

anywhere LOG level warning ip-options prefix `WRONG IFACE:
'

(Don't look at counters; right now, for getting good internet access, I'm not using multihop)

So, often appears in the kernel log, specially with 'ftp' and 'ssh' connections (and rarely with www connections):

Jul  4 15:50:14 thecrow WRONG IFACE: IN= OUT=eth2 SRC=192.168.16.1

DST=216.165.191.52 LE
N=72 TOS=0x00 PREC=0x00 TTL=64 ID=9582 DF PROTO=TCP SPT=56528 DPT=6667 WINDOW=18824 RES=
0x00 ACK PSH URGP=0

Jul  4 16:01:29 thecrow WRONG IFACE: IN= OUT=eth1 SRC=192.168.17.1

DST=130.206.1.5 LEN=40 TOS=0x00 PREC=0x00 TTL=64 ID=66 DF PROTO=TCP SPT=33820 DPT=21 WINDOW=0 RES=0x00 RST URGP=0

Even though, when I use 'tcpdump' for catching the wrong packets (that
is: tcpdump -i eth1 host 192.168.17.1 _or_ tcpdump -i eth2 host 192.168.16.1 ) results that _A LOT MORE PACKETS_ are BADLY ROUTED, than sent to the LOG target.

My conclusion: iptables 'matching' doesn't work; also does the route decision part.

I absolutely don't know what more to do... I'm running iptables v1.2.11, and kernel 2.6.11-gentoo-r11. Exactly same happened with kernel 2.4.28-gentoo. :(

For example, here is a test. I want to ftp to "ftp.rediris.es". I look which would be the route:

# ip route get 130.206.1.5
130.206.1.5 via 192.168.17.2 dev eth2  src 192.168.17.1 
    cache  mtu 1500 advmss 1460 metric10 64
I try the ftp:
# ftp ftp.rediris.es
And in the kernel log appears:
Jul  4 16:19:25 thecrow WRONG IFACE: IN= OUT=eth1 SRC=192.168.17.1
DST=130.206.1.5 LEN=60 TOS=0x00 PREC=0x00 TTL=64 ID=43245 DF PROTO=TCP
SPT=49828 DPT=21 WINDOW=5840 RES=0x00 SYN URGP=0

Please, help... I'm desperate.

#3 lore raw
Quoted 4 lines · Lluis Batle

I have many problems getting this thing to work. There's a host with two network interfaces, where there are two routers to Internet in two separated networks. The host uses multihop routing for deciding to which router send the packets... but the routing decision is wrong

Does your kernel have the routes patch by Julian Anastasov?
http://www.ssi.bg/~ja/

Quoted 1 line

Please, help... I'm desperate.

You'll be more likely to get help on the LARTC mailing list.

Why RFC 1918 addresses on these interfaces? Do you have extra router equipment between you and the Internet?

Are multiple "ip route get ip.add.re.ss" commands in sequence showing routes out the same interface?

Signature · 2 lines
    mail to this address is discarded unless "/dev/rob0"
    or "not-spam" is in Subject: header
#4 lore raw
Quoted 2 lines · On Monday 04 July 2005 10:07, I wrote:

Are multiple "ip route get ip.add.re.ss" commands in sequence showing routes out the same interface?

for X in seq 10 ; do /sbin/ip route get $X.$X.$X.$X ; done

Signature · 2 lines
    mail to this address is discarded unless "/dev/rob0"
    or "not-spam" is in Subject: header
#6 lore raw
In reply to a message not in the archive

No, everything is right in its output:
1.1.1.1 via 192.168.16.2 dev eth1 src 192.168.16.1

    cache  mtu 1500 advmss 1460 metric10 64
2.2.2.2 via 192.168.17.2 dev eth2  src 192.168.17.1
    cache  mtu 1500 advmss 1460 metric10 64
3.3.3.3 via 192.168.16.2 dev eth1  src 192.168.16.1
    cache  mtu 1500 advmss 1460 metric10 64
4.4.4.4 via 192.168.17.2 dev eth2  src 192.168.17.1
    cache  mtu 1500 advmss 1460 metric10 64
5.5.5.5 via 192.168.16.2 dev eth1  src 192.168.16.1
    cache  mtu 1500 advmss 1460 metric10 64
6.6.6.6 via 192.168.17.2 dev eth2  src 192.168.17.1
    cache  mtu 1500 advmss 1460 metric10 64
7.7.7.7 via 192.168.16.2 dev eth1  src 192.168.16.1
    cache  mtu 1500 advmss 1460 metric10 64
8.8.8.8 via 192.168.17.2 dev eth2  src 192.168.17.1
    cache  mtu 1500 advmss 1460 metric10 64
9.9.9.9 via 192.168.16.2 dev eth1  src 192.168.16.1
    cache  mtu 1500 advmss 1460 metric10 64
10.10.10.10 via 192.168.17.2 dev eth2  src 192.168.17.1
    cache  mtu 1500 advmss 1460 metric10 64

About the 16.x and 17.x addresses... yes, there are other routers, which make NAT (192.168.16.2 and 192.168.17.2) to internet.

Quoted 10 lines · On 7/4/05, /dev/rob0 <[email protected]> wrote:
Quoted 2 lines · On Monday 04 July 2005 10:07, I wrote:

Are multiple "ip route get ip.add.re.ss" commands in sequence showing routes out the same interface?

for X in seq 10 ; do /sbin/ip route get $X.$X.$X.$X ; done --
mail to this address is discarded unless "/dev/rob0" or "not-spam" is in Subject: header

#7 lore raw
Quoted 6 lines · On 7/4/05, /dev/rob0 <[email protected]> wrote:
Quoted 5 lines · On Monday 04 July 2005 10:07, I wrote:
Quoted 2 lines

Are multiple "ip route get ip.add.re.ss" commands in sequence showing routes out the same interface?

for X in seq 10 ; do /sbin/ip route get $X.$X.$X.$X ; done

Quoted 10 lines · Lluis Batle

No, everything is right in its output:
1.1.1.1 via 192.168.16.2 dev eth1 src 192.168.16.1

    cache  mtu 1500 advmss 1460 metric10 64
2.2.2.2 via 192.168.17.2 dev eth2  src 192.168.17.1
    cache  mtu 1500 advmss 1460 metric10 64
Quoted 4 lines

="masquerading.multi-eth" (misnamed: it does no masquerading)

NE1=192.168.16.0/28
NE2=192.168.17.0/28

Let's see, those are .0-.15 on the last quad.

Quoted 1 line
Quoted 1 line

NLOCAL=192.168.0.0/20

And this is 0.0 through 15.255 ... IOW, wrong, excluding both $NE1 and $NE2. Try 192.168.16.0/23. It would not hurt for you to brush up on TCP/IP and subnetting basics.

Quoted 10 lines
Quoted 10 lines

$IPTABLES -t nat -F PREROUTING
$IPTABLES -t nat -F POSTROUTING
$IPTABLES -t nat -F OUTPUT
$IPTABLES -t filter -F INPUT
$IPTABLES -t filter -F FORWARD
$IPTABLES -t filter -F OUTPUT
$IPTABLES -t filter -F keep_state >&/dev/null
$IPTABLES -t filter -X keep_state >&/dev/null
$IPTABLES -t nat -F keep_state >&/dev/null
$IPTABLES -t nat -X keep_state >&/dev/null

Could be rewritten as:
iptables -F ; iptables -X ; iptables -t nat -F ; iptables -t nat -X

Quoted 9 lines
Quoted 9 lines

$IPTABLES -t filter -N keep_state
$IPTABLES -t filter -A keep_state -m state \
--state RELATED,ESTABLISHED -j ACCEPT
$IPTABLES -t filter -A keep_state -j RETURN

$IPTABLES -t nat -N keep_state
$IPTABLES -t nat -A keep_state -m state \
--state RELATED,ESTABLISHED -j ACCEPT
$IPTABLES -t nat -A keep_state -j RETURN

1. IMO it's confusing to give chains the same name in different tables.
2. The RETURN rules are pointless. That's what happens at the end of a chain, anyway.
3. --state in -t nat? Is that possible? Does it work? Does it break anything?

Quoted 2 lines

About the 16.x and 17.x addresses... yes, there are other routers, which make NAT (192.168.16.2 and 192.168.17.2) to internet.

This seems odd to me. I prefer to use external IP directly, for many reasons. It also eliminates other potential points of failure.

It's even more odd considering that you're doing DNAT on the already- NAT'ed Linux machine. Why not do the DNAT in the external routers? Also, those DNAT rules refer to other RFC 1918 netblocks.

Signature · 2 lines
    mail to this address is discarded unless "/dev/rob0"
    or "not-spam" is in Subject: header
#8 lore raw

Thanks :) I answer between lines...

Quoted 1 line · On 7/4/05, /dev/rob0 <[email protected]> wrote:
Quoted 1 line
Quoted 1 line

="masquerading.multi-eth" (misnamed: it does no masquerading)

Ok. I tried with MASQUERADE, but by now I use SNAT.

Quoted 11 lines
Quoted 3 lines
Quoted 3 lines

NE1=192.168.16.0/28
NE2=192.168.17.0/28

Let's see, those are .0-.15 on the last quad.

Quoted 1 line
Quoted 1 line

NLOCAL=192.168.0.0/20

And this is 0.0 through 15.255 ... IOW, wrong, excluding both $NE1 and $NE2. Try 192.168.16.0/23. It would not hurt for you to brush up on TCP/IP and subnetting basics.

Oh. Is it wrong? I don't understand what's "IOW". Where should I try your proposed subnet? why?

Quoted 12 lines
Quoted 11 lines
Quoted 11 lines

$IPTABLES -t nat -F PREROUTING
$IPTABLES -t nat -F POSTROUTING
$IPTABLES -t nat -F OUTPUT
$IPTABLES -t filter -F INPUT
$IPTABLES -t filter -F FORWARD
$IPTABLES -t filter -F OUTPUT
$IPTABLES -t filter -F keep_state >&/dev/null
$IPTABLES -t filter -X keep_state >&/dev/null
$IPTABLES -t nat -F keep_state >&/dev/null
$IPTABLES -t nat -X keep_state >&/dev/null

Quoted 2 lines

Could be rewritten as:
iptables -F ; iptables -X ; iptables -t nat -F ; iptables -t nat -X

Ok :)

Quoted 6 lines
Quoted 6 lines
Quoted 6 lines

$IPTABLES -t filter -N keep_state
$IPTABLES -t filter -A keep_state -m state \
--state RELATED,ESTABLISHED -j ACCEPT
$IPTABLES -t filter -A keep_state -j RETURN

Quoted 6 lines
Quoted 4 lines
Quoted 4 lines

$IPTABLES -t nat -N keep_state
$IPTABLES -t nat -A keep_state -m state \
--state RELATED,ESTABLISHED -j ACCEPT
$IPTABLES -t nat -A keep_state -j RETURN

1. IMO it's confusing to give chains the same name in different tables.

I agree... but by now does that matter?

Quoted 2 lines

2. The RETURN rules are pointless. That's what happens at the end of a chain, anyway.

Ok. That's what I thought

Quoted 2 lines

3. --state in -t nat? Is that possible? Does it work? Does it break anything?

It seems it's possible. I get no error from those commands. Anyway, I've thought that happens double application of that rule, through filter and nat tables. I've removed everything about 'keep_state' in the nat table. Everything is still working bad. Even from the computer itself (local processes). Routing doesn't work, even without any "NAT"-related chain/rule.

Quoted 6 lines
Quoted 3 lines

About the 16.x and 17.x addresses... yes, there are other routers, which make NAT (192.168.16.2 and 192.168.17.2) to internet.

This seems odd to me. I prefer to use external IP directly, for many reasons. It also eliminates other potential points of failure.

I agree :) but it's hard for me to configure the routers, so they give to me the public addresses. I think that double-NAT should not break anything. :)

Quoted 4 lines

It's even more odd considering that you're doing DNAT on the already- NAT'ed Linux machine. Why not do the DNAT in the external routers? Also, those DNAT rules refer to other RFC 1918 netblocks.

mmm I've never read RFC 1918. :) I'll take a look at it. Thanks!

- I finish the working time in few minutes. Tomorrow morning I'll keep on with this.

#9 lore raw
Quoted 1 line · Lluís Batlle

Thanks :) I answer between lines...

Thank you.

Quoted 4 lines · On 7/4/05, /dev/rob0 <[email protected]> wrote:
Quoted 1 line
Quoted 1 line
Quoted 1 line

="masquerading.multi-eth" (misnamed: it does no masquerading)

Ok. I tried with MASQUERADE, but by now I use SNAT.

Right. MASQUERADE will not work with multiple routing.

Quoted 5 lines
Quoted 5 lines
Quoted 2 lines
Quoted 2 lines

NE1=192.168.16.0/28
NE2=192.168.17.0/28

Let's see, those are .0-.15 on the last quad.

Quoted 8 lines
Quoted 5 lines
Quoted 1 line
Quoted 1 line

NLOCAL=192.168.0.0/20

And this is 0.0 through 15.255 ... IOW, wrong, excluding both $NE1 and $NE2. Try 192.168.16.0/23. It would not hurt for you to brush up on TCP/IP and subnetting basics.

Oh. Is it wrong? I don't understand what's "IOW". Where should I try your proposed subnet? why?

IOW="in other words", a common Internet shorthand.

192.168.0.0/20, set as $NLOCAL in your iptables script, excludes your IP addresses and networks. No packet hitting the rules which refer to that value will match, so the rules are ignored.

The rules to which I am referring:
$IPTABLES -t nat -A POSTROUTING -o eth1 -s $NLOCAL -j SNAT --to $IPE1 $IPTABLES -t nat -A POSTROUTING -o eth2 -s $NLOCAL -j SNAT --to $IPE2 Your SNAT rules.

Change "NLOCAL=192.168.0.0/20" to "NLOCAL=192.168.0.0/16", or as previously suggested, "NLOCAL=192.168.16.0/23". I suppose you could even omit the source specification altogether:
$IPTABLES -t nat -A POSTROUTING -o eth1 -j SNAT --to $IPE1 $IPTABLES -t nat -A POSTROUTING -o eth2 -j SNAT --to $IPE2

Quoted 4 lines
Quoted 2 lines

1. IMO it's confusing to give chains the same name in different tables.

I agree... but by now does that matter?

Simply a point of style. You can give chains any names you wish, no matter how confusing they might be in context:

### Kids, don't try this at home. Professional stunt driver on a ### closed track.
iptables -N InputLogDrop
iptables -N ForwardAllow
iptables -A InputLogDrop -j ACCEPT
iptables -A FORWARD -j InputLogDrop
iptables -A ForwardAllow -j LOG
iptables -A ForwardAllow -p tcp -j REJECT
iptables -A ForwardAllow -j DROP
iptables -A INPUT -j ForwardAllow
### For my next trick, I will campaign to be elected Prime Minister. ### Thank you for your support in the polls.

Quoted 4 lines
Quoted 2 lines

3. --state in -t nat? Is that possible? Does it work? Does it break anything?

It seems it's possible. I get no error from those commands. Anyway,

Perhaps it doesn't break anything, but I have read here that only packets of --state NEW hit the -t nat PREROUTING chain. I don't know about the relationship between connection tracking and NAT.

Quoted 3 lines

I've thought that happens double application of that rule, through filter and nat tables. I've removed everything about 'keep_state' in the nat table. Everything is still working bad. Even from the

Likely because of NLOCAL in your script. If that's not the case it's beyond my limited understanding, and once again I'll suggest you take it to LARTC. Some people in LARTC know more about this than I do.

Quoted 4 lines
Quoted 2 lines

already-NAT'ed Linux machine. Why not do the DNAT in the external routers? Also, those DNAT rules refer to other RFC 1918 netblocks.

mmm I've never read RFC 1918. :) I'll take a look at it.

"RFC 1918 netblocks" is simply another form of shorthand to refer to IPv4 ranges which are reserved for private use, namely 10.0.0.0/8, 172.16.0.0/12, and 192.168.0.0/16. I rarely read RFC's myself (but I must confess to a fondness for RFC 1149. :) )

Signature · 2 lines
    mail to this address is discarded unless "/dev/rob0"
    or "not-spam" is in Subject: header
#10 lore raw

Again between lines...

Quoted 12 lines · On 7/4/05, /dev/rob0 <[email protected]> wrote:
Quoted 11 lines · Lluís Batlle
Quoted 10 lines
Quoted 2 lines
Quoted 2 lines

NE1=192.168.16.0/28
NE2=192.168.17.0/28

Let's see, those are .0-.15 on the last quad.

Quoted 1 line
Quoted 1 line

NLOCAL=192.168.0.0/20

And this is 0.0 through 15.255 ... IOW, wrong, excluding both $NE1 and $NE2. Try 192.168.16.0/23. It would not hurt for you to brush up on TCP/IP and subnetting basics.

Quoted 8 lines
Quoted 2 lines

Oh. Is it wrong? I don't understand what's "IOW". Where should I try your proposed subnet? why?

IOW="in other words", a common Internet shorthand.

192.168.0.0/20, set as $NLOCAL in your iptables script, excludes your IP addresses and networks. No packet hitting the rules which refer to that value will match, so the rules are ignored.

Why? in the LAN (eth0, 192.168.0.0/20) there are many computers... if I change it to 192.168.0.0/16, eth1 and eth2 _won't_ be appart subnetworks! It's important to them to be excluded.
IOW, there must be no intersection between the networks of the different NICs.

Quoted 11 lines

The rules to which I am referring:
$IPTABLES -t nat -A POSTROUTING -o eth1 -s $NLOCAL -j SNAT --to $IPE1 $IPTABLES -t nat -A POSTROUTING -o eth2 -s $NLOCAL -j SNAT --to $IPE2 Your SNAT rules.

Change "NLOCAL=192.168.0.0/20" to "NLOCAL=192.168.0.0/16", or as previously suggested, "NLOCAL=192.168.16.0/23". I suppose you could even omit the source specification altogether:
$IPTABLES -t nat -A POSTROUTING -o eth1 -j SNAT --to $IPE1 $IPTABLES -t nat -A POSTROUTING -o eth2 -j SNAT --to $IPE2

Will, that way, the kernel maintain connection-tables for SNAT even for local connections?

Quoted 4 lines

### Kids, don't try this at home. Professional stunt driver on a ### closed track.
iptables -N InputLogDrop
iptables -N ForwardAllow

Quoted 8 lines

iptables -A InputLogDrop -j ACCEPT
iptables -A FORWARD -j InputLogDrop
iptables -A ForwardAllow -j LOG
iptables -A ForwardAllow -p tcp -j REJECT
iptables -A ForwardAllow -j DROP
iptables -A INPUT -j ForwardAllow
### For my next trick, I will campaign to be elected Prime Minister. ### Thank you for your support in the polls.

:)))

Quoted 3 lines

Perhaps it doesn't break anything, but I have read here that only packets of --state NEW hit the -t nat PREROUTING chain. I don't know about the relationship between connection tracking and NAT.

Can you give a link about that?

Quoted 5 lines

"RFC 1918 netblocks" is simply another form of shorthand to refer to IPv4 ranges which are reserved for private use, namely 10.0.0.0/8, 172.16.0.0/12, and 192.168.0.0/16. I rarely read RFC's myself (but I must confess to a fondness for RFC 1149. :) )

Hahaha :)

#11 lore raw

Re: Problem with routing decisions, and multihop (solved)

Argh. Stupid me..............
The tables 201 and 202, who decide the route for packets with already defined source addres were in the rules that way:
The routing rules are:

0:      from all lookup local
50:     from all lookup main
201:    from 192.168.17.0/28 iif eth2 lookup 201
202:    from 192.168.16.0/28 iif eth1 lookup 202
222:    from all lookup 222
32766:  from all lookup main
32767:  from all lookup default

Look at the terrible "iif ethx"!!!!!!!!! The packets were routed _at random_ by the 'equalizing routing decision', even if with decided address. Argh.
Solution: there should not be "if ethx" in the routing rule matching

Now it works.

Even though, I'd prefer to get the Julian's patches working. When I try them, all packets go to the same interface (the first hop). I'll keep on working...

Always, always, there is the deadly "user fault". :)))

Thanks for your support!