Open-Mesh » batman-adv

In a certain setup the primary interface OGMs are missing on a node's secondary interface. It might have to do with the actual first interface being down. See the following description how this has been tracked down. It also seems to be the case after every reboot.

Node A:

[B.A.T.M.A.N. adv 2011.1.0, MainIF/MAC: wlan3/0a:1b:b1:f2:f3:37 (bat0)]
  Originator      last-seen (#/255)           Nexthop [outgoingIF]:   Potential nexthops ...
00:22:b0:98:94:06    0.570s   (225) 7e:19:6a:b8:d5:d9 [     3micc]: 0a:22:b0:98:85:5b (  0) 7e:19:6a:b8:d5:d9 (225)
8e:3d:c2:10:10:28    0.150s   (241) 7e:19:6a:b8:d5:d9 [     3micc]: 7e:19:6a:b8:d5:d9 (241)
7e:19:6a:b8:d5:d9    0.080s   (251) 7e:19:6a:b8:d5:d9 [     3micc]: 0a:22:b0:98:85:5b (  0) 7e:19:6a:b8:d5:d9 (251)
00:22:b0:98:85:5c    0.070s   (249) 0a:22:b0:98:85:5b [     wlan3]: 0a:22:b0:98:86:53 (  0) 0a:22:b0:98:85:5b (249)
0a:22:b0:98:85:5b    0.090s   (248) 0a:22:b0:98:85:5b [     wlan3]: 0a:22:b0:98:86:53 (234) 0a:22:b0:98:85:5b (248)
0a:22:b0:98:86:53    0.590s   (240) 0a:22:b0:98:86:53 [     wlan3]: 0a:22:b0:98:85:5b (225) 0a:22:b0:98:86:53 (240)
00:22:b0:98:86:54    0.540s   (241) 0a:22:b0:98:86:53 [     wlan3]: 0a:22:b0:98:86:53 (241) 0a:22:b0:98:85:5b (  0)
da:7b:6f:c1:63:d2    0.960s   (227) 7e:19:6a:b8:d5:d9 [     3micc]: 7e:19:6a:b8:d5:d9 (227)

=> pimary interface address: 0a:1b:b1:f2:f3:37 and sees the following node B:

Node B:

[B.A.T.M.A.N. adv 2011.0.0 v2011.0.0-dirty, MainIF/MAC: 3micc/7e:19:6a:b8:d5:d9 (bat0)]
  Originator      last-seen (#/255)           Nexthop [outgoingIF]:   Potential nexthops ...
46:b5:37:88:46:48    0.696s   (255) 46:b5:37:88:46:48 [     3micc]: 46:b5:37:88:46:48 (255)
00:22:b0:98:94:06    0.526s   (235) 8e:3d:c2:10:10:28 [     3micc]: 8e:3d:c2:10:10:28 (235)
8e:3d:c2:10:10:28    0.223s   (255) 8e:3d:c2:10:10:28 [     3micc]: 8e:3d:c2:10:10:28 (255)
00:22:b0:98:85:5c    0.166s   (245) 46:b5:37:88:46:48 [     3micc]: 46:b5:37:88:46:48 (245)
0a:22:b0:98:85:5b   46.053s   (231) 46:b5:37:88:46:48 [     3micc]: 46:b5:37:88:46:48 (231)
00:22:b0:98:86:54    0.346s   (239) 46:b5:37:88:46:48 [     3micc]: 46:b5:37:88:46:48 (239)
da:7b:6f:c1:63:d2    0.616s   (245) 8e:3d:c2:10:10:28 [     3micc]: 8e:3d:c2:10:10:28 (245)

=> Does not have node A's primary if address! However it only has 46:b5:37:88:46:48, which is...

Node A:

root@OpenWrt:~# batctl if  
eth0: inactive
wlan3: active
3micc: active

Node A:

root@OpenWrt:~# ifconfig | grep HWaddr     
3micc     Link encap:Ethernet  HWaddr 46:B5:37:88:46:48  
bat0      Link encap:Ethernet  HWaddr 72:4A:9B:CC:88:FC  
br-mesh   Link encap:Ethernet  HWaddr 00:0D:B9:20:8F:05  
eth1      Link encap:Ethernet  HWaddr 00:0D:B9:20:8F:05  
mon.wlan1 Link encap:UNSPEC  HWaddr 00-1B-B1-F2-F3-37-20-00-00-00-00-00-00-00-00-00  
wlan0     Link encap:Ethernet  HWaddr 00:1B:B1:F2:F3:E4  
wlan1     Link encap:Ethernet  HWaddr 00:1B:B1:F2:F3:37  
wlan2     Link encap:Ethernet  HWaddr 06:1B:B1:F2:F3:37  
wlan3     Link encap:Ethernet  HWaddr 0A:1B:B1:F2:F3:37

=> ... a secondary interface address of node A.

Node A:

root@OpenWrt:~# batctl td -p 1 3micc
22:06:06.573005 BAT 00:22:b0:98:85:5c: OGM via neigh 46:b5:37:88:46:48, seq 8551, tq 232, ttl 49, v 12, flags [....], length 36
22:06:06.739677 BAT 00:22:b0:98:85:5c: OGM via neigh 7e:19:6a:b8:d5:d9, seq 8551, tq 222, ttl 48, v 12, flags [....], length 36
22:06:06.851023 BAT 8e:3d:c2:10:10:28: OGM via neigh 7e:19:6a:b8:d5:d9, seq 550894, tq 245, ttl 49, v 12, flags [D..G], length 30
22:06:06.902458 BAT 46:b5:37:88:46:48: OGM via neigh 46:b5:37:88:46:48, seq 1203, tq 255, ttl  2, v 12, flags [....], length 24
22:06:06.952537 BAT 8e:3d:c2:10:10:28: OGM via neigh 46:b5:37:88:46:48, seq 550894, tq 235, ttl 48, v 12, flags [...G], length 30
22:06:07.062312 BAT 46:b5:37:88:46:48: OGM via neigh 7e:19:6a:b8:d5:d9, seq 1203, tq 245, ttl  1, v 12, flags [D...], length 24
22:06:07.301592 BAT 00:22:b0:98:94:06: OGM via neigh 7e:19:6a:b8:d5:d9, seq 638769, tq 225, ttl 47, v 12, flags [....], length 72
22:06:07.443968 BAT 7e:19:6a:b8:d5:d9: OGM via neigh 7e:19:6a:b8:d5:d9, seq 497204, tq 255, ttl 50, v 12, flags [..F.], length 30
22:06:07.551531 BAT 7e:19:6a:b8:d5:d9: OGM via neigh 46:b5:37:88:46:48, seq 497204, tq 245, ttl 49, v 12, flags [D...], length 30
22:06:07.571635 BAT 00:22:b0:98:85:5c: OGM via neigh 46:b5:37:88:46:48, seq 8552, tq 228, ttl 49, v 12, flags [....], length 36
22:06:07.730785 BAT 00:22:b0:98:85:5c: OGM via neigh 7e:19:6a:b8:d5:d9, seq 8552, tq 219, ttl 48, v 12, flags [....], length 36
22:06:07.839376 BAT 8e:3d:c2:10:10:28: OGM via neigh 7e:19:6a:b8:d5:d9, seq 550895, tq 245, ttl 49, v 12, flags [D..G], length 30
22:06:07.881495 BAT 46:b5:37:88:46:48: OGM via neigh 46:b5:37:88:46:48, seq 1204, tq 255, ttl  2, v 12, flags [....], length 24
22:06:07.941561 BAT 8e:3d:c2:10:10:28: OGM via neigh 46:b5:37:88:46:48, seq 550895, tq 235, ttl 48, v 12, flags [...G], length 30
22:06:08.046184 BAT 46:b5:37:88:46:48: OGM via neigh 7e:19:6a:b8:d5:d9, seq 1204, tq 245, ttl  1, v 12, flags [D...], length 24

=> Which does not list 0a:1b:b1:f2:f3:37 but only 46:b5:37:88:46:48

Node A's wlan3, the primary interface seems up and running fine though:

root@OpenWrt:~# ifconfig wlan3
wlan3     Link encap:Ethernet  HWaddr 0A:1B:B1:F2:F3:37
          inet6 addr: fe80::81b:b1ff:fef2:f337/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1527  Metric:1
          RX packets:18731 errors:0 dropped:0 overruns:0 frame:0
          TX packets:9683 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:1113162 (1.0 MiB)  TX bytes:760959 (743.1 KiB)

dmesg output on node A:

root@OpenWrt:~# dmesg | grep batman
batman_adv: B.A.T.M.A.N. advanced 2011.1.0 (compatibility version 12) loaded
batman_adv: bat0: Adding interface: eth0
batman_adv: bat0: The MTU of interface eth0 is too small (1500) to handle the transport of batman-adv packets. Packets going over this interface will be fra.
batman_adv: bat0: Not using interface eth0 (retrying later): interface not active
batman_adv: bat0: Adding interface: wlan3
batman_adv: bat0: Interface activated: wlan3
batman_adv: bat0: Adding interface: 3micc
batman_adv: bat0: The MTU of interface 3micc is too small (1280) to handle the transport of batman-adv packets. Packets going over this interface will be fr.
batman_adv: bat0: Interface activated: 3micc
batman_adv: bat0: Changing gw mode from: off to: client

Note: On node A's primary interface, wlan3, the primary interface OGMs can be seen though:

root@OpenWrt:~# batctl td -p 1 wlan3 | grep 0a:1b:b1:f2:f3:37
22:32:43.866469 BAT 0a:1b:b1:f2:f3:37: OGM via neigh 0a:1b:b1:f2:f3:37, seq 134, tq 255, ttl 50, v 12, flags [..F.], length 42
22:32:43.975737 BAT 0a:1b:b1:f2:f3:37: OGM via neigh 0a:22:b0:98:86:53, seq 134, tq 226, ttl 48, v 12, flags [D...], length 42
22:32:43.978377 BAT 0a:1b:b1:f2:f3:37: OGM via neigh 0a:22:b0:98:85:5b, seq 134, tq 240, ttl 49, v 12, flags [D...], length 42
22:32:44.277310 BAT 7e:19:6a:b8:d5:d9: OGM via neigh 0a:1b:b1:f2:f3:37, seq 498809, tq 241, ttl 49, v 12, flags [....], length 30
22:32:44.297576 BAT 00:22:b0:98:86:54: OGM via neigh 0a:1b:b1:f2:f3:37, seq 10140, tq 245, ttl 49, v 12, flags [....], length 42
22:32:44.437389 BAT 00:22:b0:98:85:5c: OGM via neigh 0a:1b:b1:f2:f3:37, seq 10165, tq 235, ttl 48, v 12, flags [....], length 108
22:32:44.476302 BAT 0a:22:b0:98:86:53: OGM via neigh 0a:1b:b1:f2:f3:37, seq 10139, tq 245, ttl  1, v 12, flags [D...], length 24

Enabling eth0 and rebooting results in the following dmesg output and makes node A reachable again:

root@OpenWrt:~# dmesg | grep batman    
batman_adv: B.A.T.M.A.N. advanced 2011.1.0 (compatibility version 12) loaded
batman_adv: bat0: Adding interface: eth0
batman_adv: bat0: The MTU of interface eth0 is too small (1500) to handle the transport of batman-adv packets. Packets going over this interface will be fra.
batman_adv: bat0: Interface activated: eth0
batman_adv: bat0: Adding interface: wlan3
batman_adv: bat0: Interface activated: wlan3
batman_adv: bat0: Adding interface: 3micc
batman_adv: bat0: The MTU of interface 3micc is too small (1280) to handle the transport of batman-adv packets. Packets going over this interface will be fr.
batman_adv: bat0: Interface activated: 3micc
batman_adv: bat0: Changing gw mode from: off to: client

With node A's eth0 interface being the primary one:

root@OpenWrt:~# batctl o | head -n1
[B.A.T.M.A.N. adv 2011.1.0, MainIF/MAC: eth0/00:0d:b9:20:8f:04 (bat0)]

node B sees node A's primary interface OGMs again:

% sudo batctl o | grep 00:0d:b9:20:8f:04
00:0d:b9:20:8f:04    0.856s   (255) 16:51:9e:2b:7e:c5 [     3micc]: 16:51:9e:2b:7e:c5 (255)

=====

Node A:

root@OpenWrt:~# uname -a
Linux OpenWrt 2.6.32.27 #9 Thu May 26 21:47:19 CEST 2011 i586 GNU/Linux
root@OpenWrt:~# dmesg | grep advanced
batman_adv: B.A.T.M.A.N. advanced 2011.1.0 (compatibility version 12) loaded

Node B:

% uname -a
Linux aftermath 2.6.38.2-aftermath #1 SMP PREEMPT Tue Mar 29 03:03:35 CEST 2011 x86_64 QEMU Virtual CPU version 0.10.50 AuthenticAMD GNU/Linux
% dmesg | grep advanced
batman_adv: B.A.T.M.A.N. advanced 2011.0.0 v2011.0.0-dirty (compatibility version 12) loaded