IP connectivity issue at openwrt nodes

IP connectivity issue at openwrt nodes

[soanican]

Hi,
Im new to batman-adv and this is my first attempt to make it work. Im having connectivity issue, couldnt complete the configuration.

topology: PC1---eth0(Openwrt1)phy1sta0----------phy1ap0(openwrt2)eth0--PC2

openwrt1 has bridge(br-lan), with members of eth0 and bat0.
PC1 is connected to eth0
bat0 has interface phy1-ap0 and interface is active
openwrt1 br1 ip is 192.168.1.11

openwrt2 has bridge(br-lan), with members of eth0 and bat0
PC2 is connected to eth0
bat0 has interface phy1-sta0 and interface is active
openwrt2 br1 ip is 192.168.1.12

openwrt2 phy1-sta0 is connected to openwrt1 phy1-ap0

I can not ping 192.168.1.11 from openwrt2 and can not 192.168.1.12 from openwrt1 (from console/ssh interfaces)
My final aim is to be able to ping from PC1 connected to eth0 of openwrt1, to openwrt1, openwrt2 and PC2 connected to eth0 of openwrt2.
What I am missing here?

some logs from batman-adv:
version: batctl 2024.3-openwrt-1 [batman-adv: 2024.3-openwrt-2]

[  188.360147] phy1-sta0: associated
[  188.379399] batman_adv: bat0: Adding interface: phy1-sta0
[  188.384849] batman_adv: bat0: The MTU of interface phy1-sta0 is too small (1532) to handle the transport of batman-adv packets. Packets going over this interface will be fragmented on layer2 which could impact the performance. Setting the MTU to 1560 would solve the problem.
[  188.409162] batman_adv: bat0: Interface activated: phy1-sta0


root@dasal-01:~# brctl show
bridge name     bridge id               STP enabled     interfaces
br-lan          7fff.16acd96f6f78       no              eth0
                                                        bat0                                                        
root@owrt-01:~# batctl if
phy1-sta0: active

Re: IP connectivity issue at openwrt nodes

[Sven Eckelmann]

[-- Attachment #1: Type: text/plain, Size: 468 bytes --]

On Monday, 18 November 2024 20:44:44 CET soanican@gmail.com wrote:
> Im new to batman-adv and this is my first attempt to make it work. Im having connectivity issue, couldnt complete the configuration.

Unfortunately, this is not enough information to answer the question. And some 
of the information is only one-sided, Here are some pointers how you might be 
able to continue:

https://www.open-mesh.org/projects/batman-adv/wiki/Troubleshooting

Kind regards,
	Sven

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 228 bytes --]

Re: IP connectivity issue at openwrt nodes

[Sven Eckelmann]

[-- Attachment #1: Type: text/plain, Size: 190 bytes --]

On Tuesday, 19 November 2024 15:17:28 CET Mu De wrote:
> Dear Sven,
> 
> I would like to ask generic system and topology questions:


Don't hijack other persons threads.

Kind regards,
	Sven

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 228 bytes --]

Re: IP connectivity issue at openwrt nodes

[soanican]

Dear Sven,

I would like to ask generic system and topology questions:

1-)Generic System Question
According to info at the "https://www.kernel.org/doc/Documentation/networking/batman-adv.txt" (referred from the troubleshooting page) , if following "batman-adv" folder do not exist, your interface might not be supported. -->"/sys/class/net/eth0/batman_adv/"

If that folder does NOT exist in a system but interfaces can be add/removed via "batctl if" command, and outputs indicates interface active, can we consider this system works properly?



2-)Generic Topology Question
Consider multiple batman nodes, each have 1 bridge interface, bridge have two ports, one "bat0" and "eth0", and IP is assigned to the bridge port.  At each node, multiple wired/wireless interfaces are assigned to "bat0" which have stable connections to other nodes bat0 interfaces.  These interfaces are indicated as "active" via batctl.
Can local process of batman nodes and external non-batman nodes connected to "eth0" port of the bridge, have IP connectivity to each other via IP addresses assigned to bridges?

best regards
MuDe

Re: IP connectivity issue at openwrt nodes

[Sven Eckelmann]

[-- Attachment #1: Type: text/plain, Size: 1862 bytes --]

On Tuesday, 19 November 2024 15:17:28 CET Mu De wrote:
> Dear Sven,
> 
> I would like to ask generic system and topology questions:

Sorry about the "hijack part". Looked at first glance like another person + 
topic. Simply because you've used a different from + signature and suddenly 
added other persons in Cc.

> 1-)Generic System Question
> According to info at the "
> https://www.kernel.org/doc/Documentation/networking/batman-adv.txt" , if
> following "batman-adv" folder do not exist, your interface might not be
> supported.
> If those folder does NOT exist in a system but interfaces can be
> add/removed via "batctl if" command, and outputs indicates interface
> active, can we consider this system works properly?
> 
> "/sys/class/net/eth0/batman_adv/"

This is an outdated document. See 
https://www.kernel.org/doc/html/latest/networking/batman-adv.html for the 
correct one

> 2-)Generic Topology Question
> Consider multiple batman nodes, each have 1 bridge interface, bridge have
> two ports, one "bat0" and "eth0",

Ok, ethernet bridge to mesh

> and IP is assigned to the bridge node.

Ok

> At
> each node, multiple wired/wireless interfaces which have stable connections
> to other nodes are assigned to "bat0" and these interfaces are indicated as
> "active" via batctl.

Ok, the underlying interfaces for bat0 have "redundancy"?

> Can local process of batman nodes and external non-batman nodes connected
> to "eth0" port of the bridge, have IP connectivity to each other via IP
> addresses assigned to bridges?

Yes. The bridge is handling this part - not batman-adv. Just make sure that 
you don't have any IP conflicts. So, for example eth0 + bat0 + underlying 
interfaces for bat0 should most likely have no IP address assigned - unless 
you know what you are doing and need it for debugging (or similar things).

Kind regards,
	Sven

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 228 bytes --]

Re: IP connectivity issue at openwrt nodes

[soanican]

Hi Sven,
Yes: "Ok, ethernet bridge to mesh"
Yes, for redundancy, enabling alternative links, expecting batman handling the best links, best routes (enabling the loop avoidance): " Ok, the underlying interfaces for bat0 have "redundancy"?"
Yes, IP is only assigned to bridge interface. Making sure that interfaces are up, ensuring layer 2 connection establishment at all related nodes.

My issue is just here: I can not ping neither from local nor eth0 PC to other nodes/external client. "batctl o" shows empty table. No logs at dmesg. All bat0 sub interfaces are "active".
Any clue for troubleshooting helps a lot.

Best regards,

Re: IP connectivity issue at openwrt nodes

[Sven Eckelmann]

[-- Attachment #1: Type: text/plain, Size: 962 bytes --]

On Tuesday, 19 November 2024 15:59:58 CET Mu De wrote:
> "batctl o" shows empty table

Which means that the underlying link is not working. Either because the 
interface is down, not connected or is manipulating the transmitted OGMs 
(broadcast) in some form. The latter can for example happen if the underlying 
link is converting broadcasts to unicast - some accesspoints are doing that. I 
just mention this because you seem to use station interfaces as underlying 
interface for bat0.

You can use tcpdump on the underlying interface to see whether the submitted 
packets (OGMs) on one side are received exactly the same on the other side. 
Wireshark has support for batman-adv. So, you could easily check the recorded 
PCAPs with it.

You can also check "batctl n" to see on which side you can correctly "hear" 
OGM packets. If you can not hear it only on one side then you should check 
this one initially (but continue with the rest).

Kind regards,
	Sven

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 228 bytes --]

Re: IP connectivity issue at openwrt nodes

[soanican]

ok, thanks for the hint.
mystery resolved:
my AP and STA were both on WDS mode, looks like batman is having issues with 4-address mac frame format.
Taking them to standard mode resolved the issue.

But then Im having many tcp out-of-order, dup-ack issues, making streaming almost impossible. Not sure if these are related to MTU stuff.
Will need to spend time on this.
If any threads regarding this may help a lot.

Best regards

Re: IP connectivity issue at openwrt nodes

[Sven Eckelmann]

[-- Attachment #1: Type: text/plain, Size: 555 bytes --]

On Tuesday, 19 November 2024 22:58:23 CET soanican@gmail.com wrote:
> my AP and STA were both on WDS mode, looks like batman is having issues with 4-address mac frame format.
> Taking them to standard mode resolved the issue.

batman-adv doesn't even see these headers. It only gets to see the normal 
2-address 802.3 header. This is why I said that you should have used 
tcpdump to capture on top of the underlying interfaces.

And to be honest, I don't get why you are using standard accesspoint 
interfaces below batman-adv anyway.

Kind regards,
	Sven

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 228 bytes --]

Re: IP connectivity issue at openwrt nodes

[soanican]

You are right about frame format. Most probably, while converting from 802.11 to 802.3, source MAC address differs so batman is not working with WDS. (can not recognize advertised interface MAC address??)

Now, packet performance  issues arise. I suspect on MTU, fragmentation related issues through cascaded/nested structure of bridge/bat0/link system. Will need to dive into monitoring/capturing (as you also mentioned) this system.
For example, streaming continues for a while and then stops. Simple restart of networking system solves the issue. If you have have a clue, it can help a lot. I will consider opening a thread for this.

Best regards,