This started out from Gradiant's big data charts but has been highly modified to work with the current version of Hive Metastore
There are no offically supported Hive Metastore images. I decided to go with the official Apache/Hive image, but just use the metastore commands. This image doesn't support S3 out of the box, so I created a simple Dockerfile which starts with the Apache image and just creates links to the S3 libraries so they are picked up by the metastore service.
Hive needs for the database and tables to exist. There is a tool built
into the docker image which can set this up. We launch the init-schema-job
to do this. It waits for Postgres to come up and builds the schema and
creates the service account. The job pod can be deleted once it reaches the
completed state.
| Repository | Name | Version |
|---|---|---|
| https://charts.bitnami.com/bitnami | postgresql | ~13.2.25 |
| Key | Type | Default | Description |
|---|---|---|---|
| conf.hiveSite.hive_metastore_uris | string | "thrift://hive-metastore:9083" |
|
| image.pullPolicy | string | "IfNotPresent" |
|
| image.repository | string | "naushadh/hive-metastore" |
|
| image.tag | string | "latest" |
|
| postgresql.postgresqlDatabase | string | "metastore" |
|
| postgresql.postgresqlPassword | string | "hive" |
|
| postgresql.postgresqlUsername | string | "hive" |
|
| resources | object | {} |
|
| objectStore.externalURI | URL | Full URL to connect to object store | |
| objectStore.accessKey | string | Access key | |
| objectStore.secretKey | string | Password to object store account |