DataStax Help Center

LCM Appears to Display Duplicate Clusters

Introduction
In certain circumstances, the OpsCenter Lifecycle Manager UI may appear to display the same cluster two times: once as a "managed" cluster configured by LCM, and once as an "unmanaged" cluster available for import.
 
Workaround for the Transient Case
In most cases, this condition is transient and is occurs when the keyspace that Opscenter uses to store data in Cassandra is temporarily unavailable. In this case, the condition can be resolved by:
  1. Ensure that the DSE cluster hosting the OpsCenter keyspace is up and running. If a storage cluster has been explicitly configured, this will be the storage cluster. If a storage cluster has not been configured, this will be the cluster being monitored.
  2. Restart OpsCenter.
  3. Ensure that OpsCenter is successfully communicating with the OpsCenter keyspace by viewing OpsCenter metrics, confirming they are updating on an ongoing basis, and waiting for any transient agent-communication or node-status warnings to clear.
  4. View the LCM interface, the duplicate cluster should no longer be visible.
 
Workaround for the Persistent Case
In rare cases, restarting OpsCenter when the OpsCenter keyspace is available is insufficient to resolve the condition. This usually occurs when the OpsCenter keyspace has been destroyed (which sometimes occurs by design during development setups). In this case, you must use the OpsCenter API to manually re-associate the LCM topology with the monitored Opscenter cluster (note that values in ALL_CAPS are site-specific and must be replaced by the relevant information from your local environment).
  1. Determine the unique-cluster-id currently assigned to the cluster by Opscenter: `curl http://127.0.0.1:8888/CLUSTER_NAME/cluster/unique_cluster_id`
  2. Get a list of LCM clusters and make note of the ID of the affected cluster. This id is different from the ids in steps 1 and 3 and is not expected to match: `curl http://127.0.0.1:8888/api/v1/lcm/clusters/`
  3. Determine the unique-cluster-id that LCM believes is assigned to the cluster by inspecting the 'opsc-cluster-id' field in the output of: `curl http://127.0.0.1:8888/api/v1/lcm/clusters/LCM-CLUSTER-UUID`
  4. Compare the OpsCenter cluster-ids that you gathered in step 1 vs step 3. If they do not match, you must update LCM with the new opsc-cluster-id as shown in step 5. If the they do match, your case is not addressed by this article and you must contact DataStax support for assistance.
  5. Update LCM with the new opsc-cluster-id: curl -X PUT http://127.0.0.1:8888/api/v1/lcm/clusters/LCM-CLUSTER-UUID -H "Content-Type: application/json" -d '{"opsc-cluster-id": "OPSC-CLUSTER-ID"}'
  6. Refresh the browser using the LCM web-interface and the duplicate cluster should no longer be visible.
 
Technical Details
OpsCenter stores a a UUID in the Opscenter keyspace. If the keyspace is unavailable when OpsCenter is starting up then LCM will temporarily be unable to associate the LCM topology with an OpsCenter cluster, and the cluster may be incorrectly displayed as available for import. When the OpsCenter keyspace is destroyed, as occasionally occurs in development environments, the association is permanently lost and must be manually restored by inspecting the UUID's and updating LCM via its API.
 
 
Was this article helpful?
0 out of 0 found this helpful
Have more questions? Submit a request

Comments

Powered by Zendesk