🚀 go-pugleaf

Hi folks,

If anyone can shed any light on this I'll be eternally grateful!

I have a cluster setup where the head node has two ethernet interfaces and
the compute node each have one. I have a license server that I need to get
the compute nodes to see which means (at first at least) setting the head
node up as a router so that the compute nodes can "see" the license server.

The head node has the following routing table:

Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
172.16.48.54    0.0.0.0         255.255.255.255 UH    0      0        0 eth0
192.168.1.0     0.0.0.0         255.255.255.0   U     0      0        0 eth1
172.16.0.0      0.0.0.0         255.255.0.0     U     0      0        0 eth0
127.0.0.0       0.0.0.0         255.0.0.0       U     0      0        0 lo
0.0.0.0         172.16.48.1     0.0.0.0         UG    0      0        0 eth0

and it's network interfaces look like this:

eth0      Link encap:Ethernet  HWaddr 00:40:F4:54:4D:12
          inet addr:172.16.48.53  Bcast:172.16.255.255  Mask:255.255.0.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:260401 errors:0 dropped:0 overruns:0 frame:0
          TX packets:15132 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:100
          RX bytes:25992199 (24.7 Mb)  TX bytes:17412457 (16.6 Mb)
          Interrupt:10 Base address:0xaf00

eth1      Link encap:Ethernet  HWaddr 00:20:ED:36:38:CE
          inet addr:192.168.1.1  Bcast:192.168.1.255  Mask:255.255.255.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:701 errors:0 dropped:0 overruns:0 frame:0
          TX packets:823 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:100
          RX bytes:80551 (78.6 Kb)  TX bytes:103916 (101.4 Kb)
          Interrupt:11 Base address:0xce00

lo        Link encap:Local Loopback
          inet addr:127.0.0.1  Mask:255.0.0.0
          UP LOOPBACK RUNNING  MTU:16436  Metric:1
          RX packets:351088 errors:0 dropped:0 overruns:0 frame:0
          TX packets:351088 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0
          RX bytes:23976015 (22.8 Mb)  TX bytes:23976015 (22.8 Mb)

The compute nodes are on the subnet 192.168.1.0/24 and the machine I  
need them to see is 172.16.48.54

I have ip forwarding switched on (/proc/sys/net/ipv4/ip_forward = 1), 
the head node can see both compute nodes and the license server, and
iptables is set to pass everything:

[root@master root]# iptables -L
Chain INPUT (policy ACCEPT)
target     prot opt source               destination

Chain FORWARD (policy ACCEPT)
target     prot opt source               destination

Chain OUTPUT (policy ACCEPT)
target     prot opt source               destination

When I'm sitting on a compute node and try to traceroute to the license
server I get this:

[root@node2 root]# traceroute cfd1
traceroute to cfd1.jhbb.com.au (172.16.48.54), 30 hops max, 38 byte packets
 1  master (192.168.1.1)  0.113 ms  0.093 ms  0.107 ms
 2  * * *
 3  * * *

But sitting on the head node (you will have guessed by now it's called
"master") I can ping both machines involved:
[root@master root]# ping node2
PING node2 (192.168.1.102) 56(84) bytes of data.
64 bytes from node2 (192.168.1.102): icmp_seq=1 ttl=64 time=0.111 ms

[root@master root]# ping cfd1
PING cfd1.jhbb.com.au (172.16.48.54) 56(84) bytes of data.
64 bytes from cfd1.jhbb.com.au (172.16.48.54): icmp_seq=1 ttl=64 time=0.128 ms

Can anyone think why packets should not be being forwarded? Anywhere else
I can look?

Thanks in advance,

Ron

PS I think I might just add a route on the license server machine so that
packets from 192.168.1.* know that they have to go back via 172.16.48.53 -
what do you think? Think that might help?