Ping a series of IPv6 targets based on some criteria, and report on the results.

So I have a bunch of hosts at a given third-party data centre that lose their IPv6 connectivity randomly. The problem is upstream (ugh) and the solution is "restart networking" (ugggh). I needed a way to test IPv6 connectivty with some flexibility - this little script scratches that particular itch.

There is an excellent IPv6 in-browser test suite called Falling Sky that powers http://test-ipv6.com, the source of which contains a list of IPv6 targets which can be pinged along with some related meta-data.

v6pinger will attempt to obtain that list, parse it according to some user-configurable options, and produce a series of hosts to test against.

There must be a file called config.yaml in the same directory as v6pinger.py. This file contains a number of configuration options, some of which are more likely to be toyed with than others - notably:

The list contains meta-data for each host that can be used to help specify which hosts should appear in the test series. For example, to ensure that any hosts provided by vr.org, or with a loc field of global or FR are included :

prefers:
    provider:
        - vr.org
    loc:
        - global
        - FR

The minimum number of targets to test. Note that v6pinger will abort if this number of targets cannot be obtained for some reason.

The maximum number of targets to test. Note that v6pinger will fill the test series with randomly-selected targets if the prefers criteria doesn't produce enough targets.

A multiplier to determine how many successful targets are required in order to declare victory. A multiplier of 0.5, for example, would mean that half of the targets must be pingable.

The maximum age (in seconds) of the cached list after which a fresh copy will be otained.

$ ./v6pinger.py

In my express case, I put it in a wrapper that's more or less like this :

#!/usr/bin/env bash

cd /opt/v6pinger/
. ./.venv/bin/activate
./v6pinger.py
if [[ $? -eq 1 ]]; then
    echo "[ERROR] Something went wrong!" 1>&2
elif [[ $? -gt 1 ]]; then
    echo "[ERROR] restarting dibbler" 1>&2
    /usr/sbin/service dibbler-client restart
fi
deactivate

With a cron job to capture STDERR as necessary :

*/5 * * * * root /opt/v6pinger/cron_wrapper.sh > /dev/null | logger -i -t v6pinger