EmptyDir has the same lifetime as the pod, whereas we want our
persistence to extend past that, e.g. when Weave Net is upgraded.
`weave reset` cleans out the file.
We assume that /var/lib is writable.
The problem is that the plugin name is prefixed with a registry name.
So, if the plugin is fetched from Docker Store, its name will be
"store/weaveworks/net-plugin".
The detection function does an exact match on "weaveworks/net-plugin",
so in the Docker Store case, the presence of the plugin cannot be
detected.
Weave Net initially had the ability to start a single component at a time with `launch-proxy` and `launch-router`.
This change enables this again via `--proxy=false` despite the merge of the proxy and the router.
Note that by default:
- `launch` passes `--proxy` which enables the proxy and keeps behaviour consistent with `launch` pre-merge,
- `weaver` defaults `--proxy` to `false` which disables the proxy and keeps behaviour consistent with `weaver` pre-merge (as it was `weaveproxy`'s responsibility).
Note: we no longer need to detect awsvpc when launching.
The "AWS VPC" mode is now set via the `--awsvpc` flag.
Before we had to "detect" it when the proxy was potentially started
later, after the router had started.
We drop WEAVEPROXY_DOCKER_ARGS as obsolete.
If this function dies in the middle, the leftover weave-kube-test set
causes this function and weave-npc to break. In particular
this means we can never restart weave with the launch script.
Here we clean up in advance, in case there's leftovers from a previous run
Fixes#2934
"docker.sock" is used to access Docker API which is required to:
1. subscribe to Docker events to configure-arp,
2. query Docker network info to detect whether multicast should be
enabled.
3. get list of Swarm peers before starting the plugin.
(1) is currently not implemented.
(2) is replaced by WEAVE_MULTICAST env var which
can enable multicast for all Weave networks created by
the plugin (not ideal, but better than nothing).
(3) is fixed by not removing "restart.sentinel".
Why:
* In some environments Nodes have an ExternalIP but no InternalIP
This change addresses the need by:
* Falling back to ExternalIP addresses if there isn't an InternalIP for a node