[Unstrung-hackers] unable to see routes with sunshine on b2b connected nodes

Vijay Anand R vjanandrgky at gmail.com
Tue Sep 22 14:39:48 EDT 2015


Thanks Michael,

curious about the command that you suggested:
/sbin/sunshine.run --dagid ripple -i eth1 --rank 1 --prefix 
2001:db8:0001::/48 --interval 30000 --verbose --timelog

so the prefix here is what is used by the nodes to auto configure 
themselves and is propagated using the DIO messages from the root to the 
rest of the DAG.
But what if I want to advertise a particular IPV6 address (/128) that is 
assigned by me ? How do I provide the address that has to be advertised 
in the DAO messages ?

Please note I am trying to configure all nodes in storing mode so I need 
to store routes and advertise DAO as well (unlike the glow program).

Regards,
Vijay

On 9/22/2015 11:09 PM, Michael Richardson wrote:
> Vijay Anand R (vanandr) <vjanandrgky at gmail.com> wrote:
>      > I am trying to evaluate unstrung implementation of RPL for one of my
>      > projects. Please note my target platform is a 64bit linux ubuntu
>      > machine, so I made some changes to the build-setup-travis.sh script and
>      > some changes to the Makefile.inc to get a clean build.
>
> okay.  Glad to have you!
>
>      > I want to advertise the IPV6 ULA addresses, and bring up the RPL daemon
>      > in storing mode on both the test beds.  Also please note I am trying to
>      > build a DODAG based on OF0 as defined by RFC 6552.
>
> [at present, Unstrung sadly lacks any metric calculations.
> It's a TBD. So OF0 vs another won't matter actually. And the trickle timers
> need some work too]
>
>      > TB-A ==== sudo sunshine --dagid fd6a:fbaa:36f9:4141:3a42:423a:3 -I 0 -i
>      > eth1 --syslog - -verbose --prefix fd6a:fbaa:36f9:0:4141:3a42:423a:3/128
>      > --rank 1
>
>      > TB-B ==== sudo sunshine --dagid fd6a:fbaa:36f9:1:4141:3a42:423a:2 -I 0
>      > -i eth1 --syslog --verbose --prefix
>      > fd6a:fbaa:36f9:0:4141:3a42:423a:2/128
>
> I think that what you've done is declare both machines as the DODAG roots
> and put them into different DODAGs.
>
> In my testing infrastructure, which is presently UML based.
> You can see a diagram at:
>      doc/testnetwork.svg
>      https://github.com/mcr/unstrung/blob/master/doc/testnetwork.svg
> (KVM can now do the things, but I don't have time yet to switch).
>
> The file
>      testing/baseconfigs/A/root/a1.sh
> is what I use on the DODAG root (A). The file testing/baseconfigs/E/root/e1.sh
> is used on a node. I mv the file because you can't "make install" on top
> of a running executable.
>
> Also, note that the dagid is considered to be a string unless it starts with
> 0x, in which case it's HEX.  yes, running it through inet_pton() to turn IPv6
> syntax into a DODAG is probably a good idea (send patch...)
>
> So, I start things with:
>
> /sbin/sunshine.run --dagid ripple -i eth1 --rank 1 --prefix 2001:db8:0001::/48 --interval 30000 --verbose --timelog
>
> /sbin/sunshine.run --dagid ripple -i eth0 -i eth1 --interval 100000 --verbose  --timelog
>
> So, I don't set the prefix on the clients; the PIO in the DIO message will
> tell the rank>0 nodes how to configure themselves.
>
> 1# sudo sunshine --dagid fd6a:fbaa:36f9:4141:3a42:423a:3 -I 0 -i eth1 \
>     --syslog --verbose --prefix fd6a:fbaa:36f9:0:4141:3a42:423a:3/128 --rank 1
> 2# sudo sunshine --dagid fd6a:fbaa:36f9:1:4141:3a42:423a:2 -I 0 -i eth1 \
>     --syslog --verbose --prefix fd6a:fbaa:36f9:0:4141:3a42:423a:2/128
>
> You seem to be setting different DAGIDs, which doesn't make any sense if you
> want them to talk to each other.  So, I think that's why they ignore each
> other.  The DAGID ought to be truncated to 16 bytes, so they actually
> should get truncated to "fd6a:fbaa:36f9:1" which is what you observed.
>
> Can you capture some logs and gist them?
>
> --
> ]               Never tell me the odds!                 | ipv6 mesh networks [
> ]   Michael Richardson, Sandelman Software Works        | network architect  [
> ]     mcr at sandelman.ca  http://www.sandelman.ca/        |   ruby on rails    [
>



More information about the Unstrung-hackers mailing list