eibkolibri: Follow channel update with channel import

If new nodes are added in a new channel version, the channel needs to be
updated or they will be skipped if they're below an already existing
node. However, if there are missing nodes from a previously failed
import, then a regular import is needed. Therefore, follow a channel
update with a channel import.

The update is run first because the algorithm for calculating newly
needed nodes depends on comparing the old and new versions of the
channel metadata. One a regular import is done, the old version of the
channel metadata is replaced and the calculation can't be made anymore.

https://phabricator.endlessm.com/T34697

lib/eibkolibri.py

@@ -190,12 +190,13 @@ def seed_channel(self, channel_id):
         """Import or update channel and content on remote Kolibri server
 
         If the channel exists, it will be updated since Kolibri won't
-        import new content nodes otherwise.
+        import new content nodes otherwise. An import is always run to
+        ensure any nodes missed because of a previous failure are
+        imported.
         """
         if self._channel_exists(channel_id):
             self.update_channel(channel_id)
-        else:
-            self.import_channel(channel_id)
+        self.import_channel(channel_id)
 
     def _get_server_series(self):
         """Determine the server Kolibri series"""