3

I have a host which is multihomed, eth0 is on 172.31.254.0/24 and eth0.10 is on 172.31.253.0/24 . Obviously eth0.10 is a subinterface on vlan id 10.

From this host I can succesfully ping hosts on the 172.31.253.0/24 network but not consistently on the 172.31.254.0/24 network. For example, notice the dropout in pings #5-#25:

[root@pbx1 ~]# ping -I eth0 172.31.254.37                      
PING 172.31.254.31 (172.31.254.31) from 172.31.254.13 eth0: 56(84) bytes of data.
64 bytes from 172.31.254.37: icmp_seq=1 ttl=128 time=1.03 ms
64 bytes from 172.31.254.37: icmp_seq=2 ttl=128 time=0.247 ms
64 bytes from 172.31.254.37: icmp_seq=3 ttl=128 time=0.236 ms
64 bytes from 172.31.254.37: icmp_seq=4 ttl=128 time=4.00 ms
64 bytes from 172.31.254.37: icmp_seq=26 ttl=128 time=0.237 ms
64 bytes from 172.31.254.37: icmp_seq=27 ttl=128 time=0.299 ms

My interfaces look right:

[root@myhost1 ~]# ifconfig
eth0      Link encap:Ethernet  HWaddr 00:22:4D:B2:28:AC  
          inet addr:172.31.254.13  Bcast:172.31.254.255  Mask:255.255.255.0
          inet6 addr: fe80::222:4dff:feb2:28ac/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:121808 errors:0 dropped:0 overruns:0 frame:0
          TX packets:120948 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:16685937 (15.9 MiB)  TX bytes:23059300 (21.9 MiB)
          Interrupt:16 Memory:d0020000-d0040000 

eth0.10 Link encap:Ethernet  HWaddr 00:22:4D:B2:28:AC  
          inet addr:172.31.253.4  Bcast:172.31.253.255  Mask:255.255.255.0
          inet6 addr: fe80::222:4dff:feb2:28ac/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:2732 errors:0 dropped:0 overruns:0 frame:0
          TX packets:828 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0 
          RX bytes:1394817 (1.3 MiB)  TX bytes:417925 (408.1 KiB)

lo        Link encap:Local Loopback  
          inet addr:127.0.0.1  Mask:255.0.0.0
          inet6 addr: ::1/128 Scope:Host
          UP LOOPBACK RUNNING  MTU:16436  Metric:1
          RX packets:30573 errors:0 dropped:0 overruns:0 frame:0
          TX packets:30573 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0 
          RX bytes:2225449 (2.1 MiB)  TX bytes:2225449 (2.1 MiB)

and the routing table looks right:

[root@myhost1 ~]# route
Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
172.31.254.0    *               255.255.255.0   U     0      0        0 eth0
172.31.253.0    *               255.255.255.0   U     0      0        0 eth0.10
link-local      *               255.255.0.0     U     1002   0        0 eth0
default         firewall.mydomain.com 0.0.0.0         UG    0      0        0 eth0

So why are my packets not getting to (or responses getting back from) hosts on the 172.31.254.0/24 network? I set /proc/sys/net/ipv4/conf/eth0/rp_filter to 0 but it makes no difference

Update: Routing table for destination host:

IPv4 Route Table
===========================================================================
Active Routes:
Network Destination        Netmask          Gateway       Interface  Metric
          0.0.0.0          0.0.0.0     172.31.254.1    172.31.254.37    266
        127.0.0.0        255.0.0.0         On-link         127.0.0.1    306
        127.0.0.1  255.255.255.255         On-link         127.0.0.1    306
  127.255.255.255  255.255.255.255         On-link         127.0.0.1    306
     172.31.252.0    255.255.255.0     172.31.254.2    172.31.254.37     11
     172.31.253.0    255.255.255.0     172.31.254.2    172.31.254.37     11
     172.31.254.0    255.255.255.0         On-link     172.31.254.37    266
    172.31.254.37  255.255.255.255         On-link     172.31.254.37    266
   172.31.254.255  255.255.255.255         On-link     172.31.254.37    266
        224.0.0.0        240.0.0.0         On-link         127.0.0.1    306
        224.0.0.0        240.0.0.0         On-link     172.31.254.37    266
  255.255.255.255  255.255.255.255         On-link         127.0.0.1    306
  255.255.255.255  255.255.255.255         On-link     172.31.254.37    266
===========================================================================
Persistent Routes:
  Network Address          Netmask  Gateway Address  Metric
     172.31.252.0    255.255.255.0     172.31.254.2       1
     172.31.253.0    255.255.255.0     172.31.254.2       1
          0.0.0.0          0.0.0.0     172.31.254.1  Default
===========================================================================
TSG
  • 1,634
  • 6
  • 29
  • 51

3 Answers3

0

You may have 172.31.254.13 used twice on your network.

I am also suspecting routing or arp issues on the receiving side:

Could you check arp -na on host 172.31.254.13 and host 172.31.254.31 ?

Full routing tables with ip route command on each host might also help.

Xavier Nicollet
  • 600
  • 3
  • 10
  • I confirmed there are no duplicate addresses (.13 or .31). The routing table for the source is already above, and I've added the routing table for the destination (a windows machine). – TSG Apr 17 '15 at 17:49
0

I needed to create an iproute2 policy, and the disable rp_filter for the interface and subinterface.

After that it worked perfectly.

TSG
  • 1,634
  • 6
  • 29
  • 51
-1

You are having a VIP on a different network on the same card. (same MAC) Try to do the same but create vNIC that are separate segments and then set a default route and specific for the other NIC.

Alex
  • 1
  • 4
  • I'm not sure I understand. The default route is already eth0 (the same subnet that I can't ping consistantly). Are you saying I should set a default route for each NIC? (I don't think that's possible) – TSG Apr 17 '15 at 17:34
  • There is only one default route. You can have routes for specific networks defined for each network. How did you create two different subnets on the same NIC? I never seen that. – Alex Apr 17 '15 at 17:47
  • Add a subinterface to a NIC to get the 2 subnets – TSG Apr 19 '15 at 21:22