[Babel-users] able to ping to neighbor but no entry in routing table

[Dave Taht]

an ip route (normal) and ip route (bad) would be helpful.

knowing what kernel versions you are running would be rather helpful. During
the routing cache elimination period (linux 3.3 through 3.8 I think), lots
of routing bugs came to light. More recently some bugs were found and fixed
in ipv6_subtrees. I'm pretty happy with babels on kernels later than 3.10.12.

On various kernels I have observed "stuck" infinite routes, bad shortest path
finding, (and on a crash of babel, incomplete removal of routes.)

A big problem is people installing default routes via methods other than babel,
(e.g. dhcp). The tyranny of the default route is a real problem,
although it is easily
disabled in dhclient (for boxes using isc-dhcp) or in openwrt (set
defaultroute 0 on all
dhcp acquired networks)

One problem with using p2p routes I've seen on ethernet is many
devices don't respond anymore to a icmp redirect, which gets issued by
some routers.

then there are inevitable collisions with various iptables rules
people seem to want to insert on routers.



On Sat, Mar 22, 2014 at 2:48 PM, Harshal Vora <harshal at amideeptech.com> wrote:
> Hi,
>
> We have setup a cluster of machines in a large room (8000 sq ft).
> One of these machines is connected to ethernet (master) and all others
> (slaves) connect to each other using adhoc wifi network.
>
> We use babeld to maintain routing tables on top of wifi adhoc network (to
> achieve mesh like functionality).
>
> We are facing an issue where all of a sudden the routing table is empty on
> the master as well as on the slave.
> We checked that we are able to ping from the master to the slave (because of
> the proximity and because they are on the same adhoc network), but we do not
> see any entry for each other in their respective routing tables.
>
> We have verified that babel daemon is running and also restarted the babel
> daemon on both the machines without any success.
>
> We run babel with logging level 3.
>
> Below are the command and logs on master as well as slave.
>
> Master: (wifi ad hoc network ip: 10.0.0.1, ethernet ip: 192.168.1.101)
>
> Command:
> start-stop-daemon --start --pidfile /var/run/babeld.pid --exec
> /usr/local/bin/babeld -- -C 'redistribute metric 128' -C 'redistribute proto
> 3 allow' -d 3 -L /var/log/babeld.log -D -I /var/run/babeld.pid -r -g 33123
> wlan0
>
> Logs:
> My id 2c:46:03:67:47:3d:6c:fd seqno 65276
> 192.168.1.101/32 metric 0 (exported)
> 10.0.0.1/32 metric 0 (exported)
> 0.0.0.0/0 metric 0 (exported)
> 10.0.0.0/24 metric 128 (exported)
> 192.168.1.0/24 metric 128 (exported)
> Sending hello 58580 (400) to wlan0.
>
>
> Slave: (ad hoc network ip: 10.0.0.9, no ethernet connection)
>
> Command:
> start-stop-daemon --start --pidfile /var/run/babeld.pid --exec
> /usr/local/bin/babeld -- -d 3 -L /var/log/babeld.log -D -I
> /var/run/babeld.pid -r -g 33123 wlan0
>
> Logs:
> My id 3c:95:86:8d:4a:4e:3f:d3 seqno 57633
> 10.0.0.9/32 metric 0 (exported)
> Sending hello 44980 (400) to wlan0
>
>
> The same logs are repeated continuously.
> Looks like it is broadcasting its own routing table.
>
> We have seen everything working properly for 1-2 hours and then suddenly
> things break. This is happening since 2 days.
> In normal working conditions, it is likely that master(10.0.0.1) connects to
> this particular slave (10.0.0.9) via another slave (10.0.0.4) in between
> these two machines which is currently down (not able to ping 10.0.0.4 from
> either of these machines).
>
> Any help will be appreciated.
> Are there any known issues with babeld for larger networks (10 machines)?
>
> Regards,
>
> _______________________________________________
> Babel-users mailing list
> Babel-users at lists.alioth.debian.org
> http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/babel-users



-- 
Dave Täht

Fixing bufferbloat with cerowrt: http://www.teklibre.com/cerowrt/subscribe.html