This error appears to take place because we're trying to load instance objects with servers, but servers are not available yet.
2014-09-10 08:26:44.084 INFO trove.cluster.service [req-687be7b0-61ca-46b0-bce0-c818431c33ea 8304a9ecfb004283a962e033ac74df6c 11d01603d47f4fe29e207e062fbe12bd] id : 'c94fdd9f-dd3e-41e0-a00d-35641aa937b0'
2014-09-10 08:26:44.098 DEBUG trove.common.strategy [req-687be7b0-61ca-46b0-bce0-c818431c33ea 8304a9ecfb004283a962e033ac74df6c 11d01603d47f4fe29e207e062fbe12bd] Loading class trove.common.strategies.mongodb.api.MongoDbAPIStrategy from (pid=17067) load_api_strategy
/opt/stack/trove/trove/common/strategy.py:29
2014-09-10 08:26:44.105 DEBUG trove.common.strategy [req-687be7b0-61ca-46b0-bce0-c818431c33ea 8304a9ecfb004283a962e033ac74df6c 11d01603d47f4fe29e207e062fbe12bd] Loading class trove.common.strategies.mongodb.api.MongoDbAPIStrategy from (pid=17067) load_api_strategy
/opt/stack/trove/trove/common/strategy.py:29
2014-09-10 08:26:44.123 ERROR trove.instance.models [req-687be7b0-61ca-46b0-bce0-c818431c33ea 8304a9ecfb004283a962e033ac74df6c 11d01603d47f4fe29e207e062fbe12bd] Could not find nova server_id(None).
2014-09-10 08:26:44.123 ERROR trove.common.exception [req-687be7b0-61ca-46b0-bce0-c818431c33ea 8304a9ecfb004283a962e033ac74df6c 11d01603d47f4fe29e207e062fbe12bd] Cannot find compute instance None for instance 11d3d56c-fb0f-4466-8fcd-3b1dd7e40835.
2014-09-10 08:26:44.123 ERROR trove.instance.models [req-687be7b0-61ca-46b0-bce0-c818431c33ea 8304a9ecfb004283a962e033ac74df6c 11d01603d47f4fe29e207e062fbe12bd] Could not load compute instance None.
2014-09-10 08:26:44.124 WARNING trove.instance.models [req-687be7b0-61ca-46b0-bce0-c818431c33ea 8304a9ecfb004283a962e033ac74df6c 11d01603d47f4fe29e207e062fbe12bd] Could not load instance 11d3d56c-fb0f-4466-8fcd-3b1dd7e40835.
2014-09-10 08:26:44.131 DEBUG trove.instance.models [req-687be7b0-61ca-46b0-bce0-c818431c33ea 8304a9ecfb004283a962e033ac74df6c 11d01603d47f4fe29e207e062fbe12bd] Instance 11d3d56c-fb0f-4466-8fcd-3b1dd7e40835 service status is new. from (pid=17067) load_instance /op
t/stack/trove/trove/instance/models.py:473
2014-09-10 08:26:44.149 ERROR trove.instance.models [req-687be7b0-61ca-46b0-bce0-c818431c33ea 8304a9ecfb004283a962e033ac74df6c 11d01603d47f4fe29e207e062fbe12bd] Could not find nova server_id(None).
2014-09-10 08:26:44.149 ERROR trove.common.exception [req-687be7b0-61ca-46b0-bce0-c818431c33ea 8304a9ecfb004283a962e033ac74df6c 11d01603d47f4fe29e207e062fbe12bd] Cannot find compute instance None for instance 7954a634-1c17-4330-8219-da9170007173.
2014-09-10 08:26:44.149 ERROR trove.instance.models [req-687be7b0-61ca-46b0-bce0-c818431c33ea 8304a9ecfb004283a962e033ac74df6c 11d01603d47f4fe29e207e062fbe12bd] Could not load compute instance None.
2014-09-10 08:26:44.150 WARNING trove.instance.models [req-687be7b0-61ca-46b0-bce0-c818431c33ea 8304a9ecfb004283a962e033ac74df6c 11d01603d47f4fe29e207e062fbe12bd] Could not load instance 7954a634-1c17-4330-8219-da9170007173.
2014-09-10 08:26:44.154 DEBUG trove.instance.models [req-687be7b0-61ca-46b0-bce0-c818431c33ea 8304a9ecfb004283a962e033ac74df6c 11d01603d47f4fe29e207e062fbe12bd] Instance 7954a634-1c17-4330-8219-da9170007173 service status is new. from (pid=17067) load_instance /op
t/stack/trove/trove/instance/models.py:473
2014-09-10 08:26:44.168 ERROR trove.instance.models [req-687be7b0-61ca-46b0-bce0-c818431c33ea 8304a9ecfb004283a962e033ac74df6c 11d01603d47f4fe29e207e062fbe12bd] Could not find nova server_id(None).
2014-09-10 08:26:44.168 ERROR trove.common.exception [req-687be7b0-61ca-46b0-bce0-c818431c33ea 8304a9ecfb004283a962e033ac74df6c 11d01603d47f4fe29e207e062fbe12bd] Cannot find compute instance None for instance 9a3d9a86-a275-438e-a207-08d7f82858a8.
2014-09-10 08:26:44.168 ERROR trove.instance.models [req-687be7b0-61ca-46b0-bce0-c818431c33ea 8304a9ecfb004283a962e033ac74df6c 11d01603d47f4fe29e207e062fbe12bd] Could not load compute instance None.
afaik, you won't see this if you use the api directly; it's because the cli actually issues a show immediately after a create. the show command for clusters does indeed attempt to load the server (as it should), so i'm not sure how we can address this outside of patching the client to stop doing this.