Re: [Babel-users] Babel for use in a dense disconnected mesh network
Are there any obvious settings that are required for babel to work in dense mesh networks? Would increasing the hello interval reduce the load on the network (though babel already appears to have a relatively low overhead). As it was noted, the issue is likely to be rooted not in babeld, but rather in lower-level wireless. One way of troubleshooting this would be replacing the Pis with an equal amount of any other PCs and confirming that the mesh builds up OK using the same USB adapters. If it goes well, the replacement may be the straightforward solution, starting from Pi-sized alternatives and going up to cheap nettops with built-in wireless, Atom CPU and no need for cross-compiling. But if it doesn't, the problem is likely in the adapters themselves or the band they are trying to use. My personal observation is that the old 802.11b is slower, yet notably more robust than the newer g/n bands. Consider borrowing a pack of different brand adapters and trying them. Eventually you may figure out specific component that fails this network. -- Denis Ovsienko ___ Babel-users mailing list Babel-users@lists.alioth.debian.org http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/babel-users
Re: [Babel-users] Babel for use in a dense disconnected mesh network
On Fri, Jun 21, 2013 at 12:26:04PM +1200, Jason Palmer wrote: Raspberry Pi, Arch Linux, babel 1.4.0, TP-Link TL-WN721N USB Wi-Fi adapter. Are you running network manager or something similar? These messages indicate that your interfaces are going up/down at a tremendous rate. I have a systemd service manually bringing up the interface, and then starting babeld. I am using the same hardware, and I have observed that the interface goes up and down spontaneously from time to time, probably because of USB rather than network issues. It might help to check that you have one of the latest Raspberry Pi (better usb design) and that your power supply is sufficient. My first tests with an early Pi and bad power supply caused the interface to go up and down several times per second (although it's probably not your case, otherwise the network would be unusable even with 10 nodes). By the way, Juliusz, I believe that ahcpd does not reassign an address to the interface when it goes up again, but I've had a hard time reproducing the issue. Is it a case it should handle correctly in theory, or a shortcoming of the implementation? Best, -- Gabriel ___ Babel-users mailing list Babel-users@lists.alioth.debian.org http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/babel-users