Bifrost Classic Support Matrix: Components and Platforms
This page lists what the Classic path migrates and what it migrates onto: the components of the default target, the supported source platforms, operating systems, JDKs, databases, and security integrations. The layout follows the support matrix published for Apache Bigtop releases.
The default target is the Ilum Bigtop distribution, managed by Apache Ambari (Ilum build), operated directly or through Ilum Cluster Ops. The default target line is certified and proven in production.
The Ilum Bigtop distribution broadly extends Apache Bigtop. It keeps services that Apache Bigtop has retired, such as Oozie, Sqoop, and Hue, and it adds services that Apache Bigtop does not package, such as Impala, Atlas, and Spark 2 alongside Spark 3. Each of them is a full management plane service with its own configuration, start order, and health checks, not a manual add-on.
Status Definitions
| Status | Meaning |
|---|---|
| Certified | Part of the default target line. Built, deployed, and validated end to end by Bifrost, including the before-and-after canary. |
| Supported | Handled by Bifrost and validated for each estate during the assessment. |
| On request | Not part of the default target line. Support is added for an engagement on request. |
Component Support
Every service below is deployed as a management plane service. Certified services are part of the default target line and are checked by the before-and-after canary; Tez is checked through Hive queries. The Roles column lists the daemons and clients that are carried. The Apache Bigtop column shows whether a service comes from Apache Bigtop or is added by the Ilum Bigtop distribution.
Storage and Resource Management
| Component | Roles | Depends on | Status | Apache Bigtop |
|---|---|---|---|---|
| एचडीएफएस | NameNode, Secondary NameNode, DataNode, JournalNode, failover controller, Router, HttpFS, client | ZooKeeper (for high availability) | Certified | Included |
| धागा | ResourceManager, NodeManager, Timeline Service, Registry DNS, client | HDFS, ZooKeeper (for high availability) | Certified | Included |
| MapReduce | JobHistory Server, client | YARN, HDFS | Certified | Included |
| ZooKeeper | Server, client | None | Certified | Included |
| Ozone | Ozone Manager, Storage Container Manager, DataNode, S3 Gateway, Recon, client | None | Supported | Added by Ilum |
SQL and Processing
| Component | Roles | Depends on | Status | Apache Bigtop |
|---|---|---|---|---|
| छत्ता | Hive Metastore, HiveServer2, WebHCat, client | HDFS, YARN, Tez, metastore database | Certified | Included |
| Tez | Client | धागा | Certified | Included |
| Impala | Statestore, Catalog Server, Impala Daemon, shell | HDFS, Hive Metastore | Certified | Added by Ilum |
| Spark 3 | History Server, client | YARN, HDFS, Hive Metastore | Certified | Included |
| Spark 2 | History Server, client | YARN, HDFS, Hive Metastore | Certified | Added by Ilum |
| Livy for Spark 3 | Livy Server | Spark 3 | Certified | Included |
| Livy for Spark 2 | Livy Server | Spark 2 | Certified | Added by Ilum |
| बहादुर | History Server, client | YARN, HDFS | Supported | Included |
Data Stores, Messaging, and Search
| Component | Roles | Depends on | Status | Apache Bigtop |
|---|---|---|---|---|
| HBase | Master, RegionServer, REST Server, Thrift Server, client | HDFS, ZooKeeper | Certified | Included |
| काफ्का | Broker | ZooKeeper | Certified | Included |
| Solr | Solr Server, for both service Solr and Infra-Solr | ZooKeeper, HDFS | Certified | Included |
| Kudu | Master, Tablet Server | Hive Metastore (for table integration) | Supported | Added by Ilum |
| Kafka Connect | Connect worker | काफ्का | Supported | Added by Ilum |
| Schema Registry | Schema Registry Server | Kafka, Schema Registry database | Supported | Added by Ilum |
| Streams Messaging Manager | SMM Server, SMM UI | Kafka, SMM database | Supported | Added by Ilum |
| Streams Replication Manager | Driver, Service | काफ्का | Supported | Added by Ilum |
| Cruise Control | Cruise Control Server | Kafka, ZooKeeper | Supported | Added by Ilum |
Security and Governance
| Component | Roles | Depends on | Status | Apache Bigtop |
|---|---|---|---|---|
| Ranger | Ranger Admin, UserSync, TagSync | Ranger database, Solr for audit | Certified | Included |
| Ranger KMS | KMS Server | Ranger, KMS database | Certified | Included |
| Atlas | Atlas Server, client | HBase, Solr, Kafka | Certified | Added by Ilum |
| Knox | Gateway | LDAP or Active Directory for authentication | Certified | Added by Ilum |
Workflow, Data Flow, and User Tools
| Component | Roles | Depends on | Status | Apache Bigtop |
|---|---|---|---|---|
| Oozie | Oozie Server, client | HDFS, YARN, Oozie database | Certified | Added by Ilum |
| Sqoop | Client | YARN, MapReduce | Certified | Added by Ilum |
| Hue | Hue Server, Kerberos Ticket Renewer | Hue database, HDFS, Hive | Certified | Added by Ilum |
| ज़ेपलिन | Zeppelin Server | Spark, Livy | Certified | Included |
| NiFi | NiFi Node, NiFi Registry | ZooKeeper (for clustering) | Supported | Added by Ilum |
Component Notes
- HDFS. NameNode high availability is carried as configured on the source. The Ilum build carries backported security fixes.
- Impala. The Ilum build of Impala carries later upstream fixes. Authorization stays with Ranger.
- Spark 3. The Apache Iceberg runtime libraries are included. The Spark Thrift Server and the
स्पार्क-एसक्यूएलshell are not shipped, matching the source. - Spark 2. Runs side by side with Spark 3. The Cloudera Spark service maps to Spark 2 and the Cloudera Spark 3 service maps to Spark 3, each with its own Livy.
- Solr. An estate with two SolrCloud instances, for example Infra-Solr and a service Solr, is given a second Solr service.
- Ranger. Plugins are connected for HDFS, YARN, Hive, HBase, Kafka, Solr, Atlas, and Knox.
- Atlas. Runs on the cluster's HBase, Solr, and Kafka. Hooks are carried for Hive, HBase, Sqoop, Impala, Kafka, and Spark.
- Zeppelin. The notebook storage format of the source is kept, so existing notebooks open unchanged.
Libraries and Agents
| Component | Used by | Apache Bigtop |
|---|---|---|
| Apache Phoenix | HBase, installed when the estate uses Phoenix | Included |
| Apache Avro, Apache Parquet, Apache ORC | Hive, Impala, Spark | Included |
| Apache Iceberg runtime | Spark 3 | Added by Ilum |
| Hive Warehouse Connector | Spark 2 and Spark 3 | Added by Ilum |
| Spark Atlas Connector | Spark 2 and Spark 3 | Added by Ilum |
| HBase Spark connector | Spark 2 and Spark 3 | Added by Ilum |
| Prometheus JMX exporter | Java daemons, for Prometheus monitoring | Added by Ilum |
Version Policy
By default, the component versions of the target are aligned with the source release. HDFS metadata, the Hive Metastore schema, client applications, and notebooks are therefore carried without an upgrade, and rollback does not have to revert data. Version differences between the source and the target are reported by discovery and reviewed in the assessment.
When an estate needs newer versions, for example a later Hadoop, Hive, or Spark release, an upgraded-version target is built for it and validated with the same canary. The versions delivered for an estate are confirmed in the assessment. See Version-aligned target.
Management Plane
| Component | Notes |
|---|---|
| Apache Ambari (Ilum build), server and agents | Bigtop stack definition extended with the services added by Ilum. Runs on JDK 17. |
| Ilum Cluster Ops | Operates the Ambari-managed cluster from Ilum. |
| Ambari alerts | Active from the first start of the adopted cluster. |
| Prometheus and Grafana | Optional, enabled with bifrost_adopt_monitoring: true. |
Apache Ambari (Ilum build) adds service definitions that the Bigtop stack of upstream Apache Ambari does not include: Atlas, Hue, Impala, Knox, Oozie, Sqoop, Spark 2 with its Livy, a second Solr, and Prometheus. See Management plane.
Source Platforms
| Source | Status | Notes |
|---|---|---|
| Cloudera CDP Private Cloud Base 7.x | Certified | The target is aligned with the source release, including the Cloudera Spark 3 parcel. |
| Cloudera CDH 6.x | Supported | The Hive Metastore schema is upgraded to the Hive 3 schema, and Sentry is translated to Ranger. |
Discovery reads Cloudera Manager 6.x and 7.x through its REST API. The API version is negotiated, and an unsupported version is refused instead of being read partially.
Operating Systems
| Operating system | Releases | Status |
|---|---|---|
| Red Hat Enterprise Linux | 8, 9, 10 | Certified |
| Rocky Linux, AlmaLinux | 8, 9, 10 | Certified |
| Oracle Linux | 8, 9, 10 | Certified |
| Ubuntu and Debian-based distributions | Current LTS and stable releases | Supported |
Rocky Linux, AlmaLinux, and Oracle Linux are binary-compatible with Red Hat Enterprise Linux and are certified on the same terms. Cluster nodes run on the x86_64 architecture; aarch64 is available on request.
The requirements for the controller host are listed in Getting started.
जावा
| JDK | Used by | Status |
|---|---|---|
| JDK 11 | All stack components (default) | Certified |
| JDK 8 | All stack components, when chosen with ambari-takeover --component-java-home | Supported |
| JDK 17 | Apache Ambari server and agents | Certified |
| JDK 17 | Stack components | On request |
One JDK is used for all adopted components. When the source runs its own JDK 11, that JDK is reused.
डेटाबेस
| डाटाबेस | हाइव मेटास्टोर | Ranger, Ranger KMS, Oozie, Hue | Apache Ambari |
|---|---|---|---|
| पोस्टग्रेएसक्यूएल | Certified | Certified | Certified (default) |
| Oracle Database | Certified | Certified | Certified |
| MySQL and MariaDB | Supported | Supported | On request |
Service databases are reused in place; no database is migrated. Apache Ambari connects to Oracle Database with the Oracle JDBC driver 19c or later.
प्रतिभूति
| Capability | Status | Notes |
|---|---|---|
| Kerberos with MIT KDC | Certified | Principals and keytabs are reused; no key version changes. |
| Kerberos with Active Directory | Certified | The KDC type is read from Cloudera Manager. |
| TLS with corporate PKI or Cloudera Manager Auto-TLS certificates | Certified | Keystores and truststores are reused, not reissued. |
| Ranger authorization | Certified | Policy content is compared before and after the cutover. |
| LDAP and Active Directory login | Certified | Atlas, Ranger Admin and UserSync, Hue, and the Ambari console. |
| HDFS encryption zones with a database-backed Ranger KMS | Certified | Every zone is verified by a decryption probe before finalize. |
| Ranger KMS backed by a hardware security module (HSM) | Certified | The HSM client is installed on the target hosts, and key custody is attested before finalize. |
| Sentry (CDH 6 sources) | Supported | Roles and grants are translated to Ranger policies. |
| Key Trustee KMS | On request | Keys are moved to Ranger KMS before the migration window. |
| Other external key stores | On request | Assessed for each estate. |
देखना Security and encryption.
Requesting Support
Support for a component, version, operating system, or integration that is not listed is added on request. A new service is delivered the same way as the services above: as packages of the Ilum Bigtop distribution, with a management plane service definition and canary checks.
To request support, contact [ईमेल संरक्षित] with the component and the version required.
अगले कदम
- Review what Classic covers and the version-aligned target.
- Read how the cluster is managed after the cutover in Management plane.
- Check how identities, certificates, and keys are carried in Security and encryption.