[Olsr-users] Output of txt_info plugin

[Dan O'Keeffe]

Hello,
I'm trying to diagnose a strange behaviour in my network using the
output from the txt_info plugin, but I'm struggling to understand how
OLSRd is coming up with its routes given the information generated by
sending an "/all" request to the txt_info plugin.


In particular, at a certain point in time, my (static) network is
working fine. In particular I'm getting good throughput between the
neighbouring nodes 10.0.0.7 and 10.0.0.3. The output of the command echo
"/all" | nc localhost:2006 on node 10.0.0.7 is below. Note especially
that in the links table, 10.0.0.3 is listed as being a symmetric link,
and as being an MPR/MPR selector. Furthermore, in the routes table
10.0.0.3 is listed as its own gateway, with an ETX of 1.301 (i.e. its
cost from the links table).

HTTP/1.0 200 OK
Content-type: text/plain

Table: Links
Local IP     Remote IP       Hyst.   LQ      NLQ     Cost
10.0.0.7     10.0.0.3        0.00    1.000   0.769   1.301
10.0.0.7     10.0.0.9        0.00    1.000   0.561   1.783

Table: Neighbors
IP address   SYM     MPR     MPRS    Will.   2 Hop Neighbors
10.0.0.9     YES     YES     YES     3       1
10.0.0.3     YES     YES     YES     3       1

Table: Topology
Dest. IP     Last hop IP     LQ      NLQ     Cost
10.0.0.4     10.0.0.2        0.709   1.000   1.408
10.0.0.6     10.0.0.2        0.674   1.000   1.482
10.0.0.5     10.0.0.3        0.674   1.000   1.482
10.0.0.2     10.0.0.4        1.000   0.678   1.474
10.0.0.8     10.0.0.4        0.983   0.769   1.321
10.0.0.3     10.0.0.5        0.983   0.674   1.506
10.0.0.8     10.0.0.5        1.000   0.662   1.509
10.0.0.2     10.0.0.6        1.000   0.674   1.482
10.0.0.9     10.0.0.6        1.000   0.748   1.335
10.0.0.9     10.0.0.7        1.000   0.561   1.783
10.0.0.4     10.0.0.8        0.772   0.983   1.314
10.0.0.5     10.0.0.8        0.662   1.000   1.509
10.0.0.6     10.0.0.9        0.753   1.000   1.328
10.0.0.7     10.0.0.9        0.607   1.000   1.645

Table: HNA
Destination  Gateway

Table: MID
IP address   Aliases

Table: Routes
Destination  Gateway IP      Metric  ETX     Interface
10.0.0.2/32  10.0.0.9        3       4.594   eth0
10.0.0.3/32  10.0.0.3        1       1.301   eth0
10.0.0.4/32  10.0.0.3        4       5.606   eth0
10.0.0.5/32  10.0.0.3        2       2.783   eth0
10.0.0.6/32  10.0.0.9        2       3.111   eth0
10.0.0.8/32  10.0.0.3        3       4.292   eth0
10.0.0.9/32  10.0.0.9        1       1.783   eth0


However, a couple of seconds later, the throughput I observe between
10.0.0.7 and 10.0.0.3 plummets. The output of txt_info now changes to
the below. The main differences I can see are that in the links table,
10.0.0.3 is no longer listed as being a symmetric link, an MPR or MPR
selector. Furthermore, the gateway for 10.0.0.3 has now become 10.0.0.9.

HTTP/1.0 200 OK
Content-type: text/plain
Table: Links
Local IP     Remote IP       Hyst.   LQ      NLQ     Cost
10.0.0.7     10.0.0.3        0.00    1.000   0.769   1.301
10.0.0.7     10.0.0.9        0.00    1.000   0.561   1.783

Table: Neighbors
IP address   SYM     MPR     MPRS    Will.   2 Hop Neighbors
10.0.0.9     YES     YES     YES     3       1
10.0.0.3     NO      NO      YES     3       1

Table: Topology
Dest. IP     Last hop IP     LQ      NLQ     Cost
10.0.0.4     10.0.0.2        0.689   1.000   1.448
10.0.0.6     10.0.0.2        0.744   1.000   1.342
10.0.0.5     10.0.0.3        0.674   1.000   1.482
10.0.0.2     10.0.0.4        1.000   0.678   1.474
10.0.0.8     10.0.0.4        0.983   0.769   1.321
10.0.0.3     10.0.0.5        0.983   0.674   1.506
10.0.0.8     10.0.0.5        1.000   0.662   1.509
10.0.0.2     10.0.0.6        1.000   0.674   1.482
10.0.0.9     10.0.0.6        1.000   0.748   1.335
10.0.0.9     10.0.0.7        1.000   0.561   1.783
10.0.0.4     10.0.0.8        0.772   0.983   1.314
10.0.0.5     10.0.0.8        0.662   1.000   1.509
10.0.0.6     10.0.0.9        0.753   1.000   1.328
10.0.0.7     10.0.0.9        0.607   1.000   1.645

Table: HNA
Destination  Gateway

Table: MID
IP address   Aliases

Table: Routes
Destination  Gateway IP      Metric  ETX     Interface
10.0.0.2/32  10.0.0.9        3       4.594   eth0
10.0.0.3/32  10.0.0.9        7       10.338  eth0
10.0.0.4/32  10.0.0.9        4       6.002   eth0
10.0.0.5/32  10.0.0.9        6       8.832   eth0
10.0.0.6/32  10.0.0.9        2       3.111   eth0
10.0.0.8/32  10.0.0.9        5       7.323   eth0
10.0.0.9/32  10.0.0.9        1       1.783   eth0

What I don't really understand is:

(i) Why the link between 10.0.0.3 and 10.0.0.7 is no longer symmetric
even though the link qualities are exactly the same as before?
(ii) Why the ETX to 10.0.0.3 in the routes tables is now 10.338, even
though its cost in the links table is only 1.301?
(iii) Why has the route to 10.0.0.3 from 10.0.0.7 changed to use
10.0.0.9 as a gateway given the topology table looks pretty much the
same as before?

Any help with the above much appreciated, even if just advice on how to
get more detailed debugging information.

Cheers,
Dan

P.s. Version/configuration information from olsrd --version is:

 *** olsr.org -  0.6.3-git_-hash_e334b612adcb7c22239a9dcafaf5b7ba  - ***
 Build date: 2013-03-28 03:47:37 on marid
 http://www.olsr.org

Parsing file: "/etc/olsrd/olsrd.conf"
Debug level: 0
        IPv4 broadcast: 255.255.255.255
        HELLO interval: 2.00
        HELLO validity: 10.00
        TC interval: 0.50
        TC validity: 30.00
        MID interval: 10.00
        MID validity: 30.00
        HNA interval: 10.00
        HNA validity: 30.00

Interface DefaultsLink quality fish eye 1
LQ Algorithm: etx_ff
IpVersion: 4
Clear screen enabled
Noint set to 1
Willingness: 3
        IPC host: 127.0.0.1
Hysteresis disabled
Link quality level 2
Pollrate 0.10
TC redundancy 2
MPR coverage 5