Jialiang Cai created AMBARI-26654:
-------------------------------------

             Summary: Add unified multi-cluster management and managed HBase 
service dependencies
                 Key: AMBARI-26654
                 URL: https://issues.apache.org/jira/browse/AMBARI-26654
             Project: Ambari
          Issue Type: New Feature
            Reporter: Jialiang Cai


Provide one Ambari server and console for multiple independent runtime 
clusters, with navigation both by cluster and by service type. Operators should 
be able to compare HBase or HDFS deployments and open the exact deployment 
configuration, operations and recovery views without repeatedly changing a 
global cluster selection.

Scope:
* Preserve Cluster -> Service -> Component -> Host and existing single-cluster 
compatibility. One host/agent belongs to one runtime cluster.
* Make cluster creation non-destructive and isolate authorization, workflow 
persistence, asynchronous requests, events, configuration, operations and 
recovery by cluster.
* Add unified React cluster/service directories, explicit cluster routes, 
useful recovery controls and improved existing installation interactions.
* Support explicit managed HBase dependencies on an HDFS service and/or an 
existing ZooKeeper service owned by another cluster, with stable identities, 
minimal versioned client configuration, private data paths, compatibility 
checks and lifecycle safeguards.
* Keep provider ownership and stored data independent of consumer stop or 
removal. Validate provider changes explicitly and preserve resumable operation 
history.
* Support the explicitly validated same-realm Kerberos contract using 
provider-local provisioning authority and consumer-only credentials. 
Unsupported authentication or migration combinations must fail clearly.

Out of scope: multiple runtime clusters sharing one host, repeated HBASE 
service identities inside one Cluster, a new Workspace tenancy model, automatic 
shared-provider deletion, and claims of process-level HA from logical 
separation.

Implementation is in progress from PR #4208's pinned source baseline 
a62fe4959dc948210d844295a097a3311321dea6. The initial contributor branch is a 
source reference checkpoint. It is not an Apache release or a completed runtime 
validation result. Compilation, focused regression tests and isolated 
cluster/browser validation remain pending; no complete test-suite result is 
claimed.

Acceptance includes non-destructive second-cluster creation and recovery, 
strict RBAC/host exclusivity, concurrent-tab/event isolation, direct 
deployment-specific actions, HBase using remote HDFS/ZooKeeper without 
redundant server deployments, safe provider impact handling and consumer 
removal preserving provider data.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to