Extracting metadata from Apache Hive for migration

This document shows how you can use the dwh-migration-dumper tool to extract the necessary metadata before running a Apache Hive data or permissions migration.

This document covers metadata extraction from the following data sources:

  • Apache Hive
  • Apache Hadoop Distributed File System (HDFS)
  • Apache Ranger
  • Cloudera Manager
  • Apache Hive query logs

Before you begin

Before you can use the dwh-migration-dumper tool, do the following:

Install Java

The server on which you plan to run dwh-migration-dumper tool must have Java 8 or higher installed. If it doesn't, download Java from the Java downloads page and install it.

Required permissions

The user account that you specify for connecting the dwh-migration-dumper tool to the source system must have permissions to read metadata from that system. Confirm that this account has appropriate role membership to query the metadata resources available for your platform. For example, INFORMATION_SCHEMA is a metadata resource that is common across several platforms.

Install the dwh-migration-dumper tool

To install the dwh-migration-dumper tool, follow these steps:

  1. On the machine where you want to run the dwh-migration-dumper tool, download the zip file from the dwh-migration-dumper tool GitHub repository.

  2. To validate the dwh-migration-dumper tool zip file, download the SHA256SUMS.txt file and run the following command:

    Bash 

    sha256sum --check SHA256SUMS.txt

    If verification fails, see Troubleshooting.

    Windows PowerShell 

    (Get-FileHash RELEASE_ZIP_FILENAME).Hash -eq ((Get-Content SHA256SUMS.txt) -Split " ")[0]

    Replace the RELEASE_ZIP_FILENAME with the downloaded zip filename of the dwh-migration-dumper command-line extraction tool release—for example, dwh-migration-tools-v1.0.52.zip

    The True result confirms successful checksum verification.

    The False result indicates verification error. Make sure the checksum and zip files are downloaded from the same release version and placed in the same directory.

  3. Extract the zip file. The extraction tool binary is in the /bin subdirectory of the folder created by extracting the zip file.

  4. Update the PATH environment variable to include the installation path for the extraction tool.

Extracting metadata for migration

Select one of the following options to learn how to extract metadata for your data source:

Apache Hive

Perform the steps in the Apache Hive section Extract metadata and query logs from your data warehouse to extract your Apache Hive metadata. You can then upload the metadata to your Cloud Storage bucket containing your migration files.

HDFS

Run the following command to extract extract metadata from HDFS using the dwh-migration-dumper tool.

dwh-migration-dumper \
  --connector hdfs \
  --host HDFS-HOST \
  --port HDFS-PORT \
  --output gs://MIGRATION-BUCKET/hdfs-dumper-output.zip \
  --assessment \

Replace the following:

  • HDFS-HOST: the HDFS NameNode hostname
  • HDFS-PORT: the HDFS NameNode port number. You can skip this argument if you are using the default 8020 port.
  • MIGRATION-BUCKET: the Cloud Storage bucket that you are using to store the migration files.

This command extracts metadata from HDFS to a file named hdfs-dumper-output.zip in the MIGRATION-BUCKET directory.

There are several known limitations when extracting metadata from HDFS:

  • Some tasks in this connector are optional and can fail, logging a full stack trade in the output. As long as the required tasks have succeeded and the hdfs-dumper-output.zip is generated, then you can proceed with the HDFS migration.
  • The extraction process might fail or run slower than expected if the configured thread pool size is too large. If you are encountering these issues, we recommend decreasing the thread pool size using the command line argument --thread-pool-size.

Apache Ranger

Run the following command to extract extract metadata from Apache Ranger using the dwh-migration-dumper tool.

dwh-migration-dumper \
  --connector ranger \
  --host RANGER-HOST \
  --port 6080 \
  --user RANGER-USER \
  --password RANGER-PASSWORD \
  --ranger-scheme RANGER-SCHEME \
  --output gs://MIGRATION-BUCKET/ranger-dumper-output.zip \
  --assessment \

Replace the following:

  • RANGER-HOST: the hostname of the Apache Ranger instance
  • RANGER-USER: the username of the Apache Ranger user
  • RANGER-PASSWORD: the password of the Apache Ranger user
  • RANGER-SCHEME: specify if Apache Ranger is using http or https. Default value is http.
  • MIGRATION-BUCKET: the Cloud Storage bucket that you are using to store the migration files.

You can also include the following optional flags:

  • --kerberos-auth-for-hadoop: replaces --user and --password, if Apache Ranger is protected by kerberos instead of basic authentication. You must run the kinit command before the dwh-migration-dumper tool tool to use this flag.
  • --ranger-disable-tls-validation: include this flag if the https certificate used by the API is self signed. For example, when using Cloudera.

This command extracts metadata from Apache Ranger to a file named ranger-dumper-output.zip in the MIGRATION-BUCKET directory.

Cloudera

Run the following command to extract metadata from Cloudera using the dwh-migration-dumper tool.

dwh-migration-dumper \
  --connector cloudera-manager \
  --url CLOUDERA-URL \
  --user CLOUDERA-USER \
  --password CLOUDERA-PASSWORD \
  --output gs://MIGRATION-BUCKET/cloudera-dumper-output.zip \
  --yarn-application-types APPLICATION-TYPES \
  --pagination-page-size PAGE-SIZE \
  --assessment \

Replace the following:

  • CLOUDERA-URL: the URL for Cloudera Manager
  • CLOUDERA-USER: the username of the Cloudera user
  • CLOUDERA-PASSWORD: the password of the Cloudera user
  • MIGRATION-BUCKET: the Cloud Storage bucket that you are using to store the migration files.
  • APPLICATION-TYPES: (Optional) list of all existing application types from Hadoop YARN. For example, SPARK, MAPREDUCE.
  • PAGE-SIZE: (Optional) specify how much data is fetched from 3rd party services, like the Hadoop YARN API. The default value is 1000, which represents 1000 entities per request.

This command extracts metadata from Cloudera to a file named dwh-migration-cloudera.zip in the MIGRATION-BUCKET directory.

Apache Hive query logs

Perform the steps in the Apache Hive section Extract query logs with the hadoop-migration-assessment logging hook to extract your Apache Hive query logs. You can then upload the logs to your Cloud Storage bucket containing your migration files.

What's next

With your extracted metadata from Hadoop, you can use these metadata files to do the following: