Created 04-16-2020 02:13 AM
Hi,
I've created hadoop cluster on Azure HDinsight, next I've created some databases on hive. After deleting and re-creating cluster with connection to the same storage account hive is not showing all databases by 'show databases;' command, I see only three default databases (default, information_schema,sys) even though the databases I've created are visible in hdfs.
Created 04-16-2020 06:35 AM
@Damian_S Yes, I always use mysql/mariadb for hive metastore. If you have the original data you can you just move it to the new location? This should be part of your migration steps regardless of the backend for the metastore.
Created on 04-16-2020 05:50 AM - edited 04-16-2020 05:51 AM
@Damian_S If you have re-created the cluster, did you migrate the original hive metastore data? This metastore would have stored details about the hive tables in hdfs. If you no longer have this metastore data, you will need to recreate the hive schemas/databases, and re-execute the hive table create statements against the hdfs files.
If the metastore is intact (not recreated, reset, etc) then you may be experiencing permission issues against the files in hdfs.
Created 04-16-2020 06:07 AM
I haven't migrated the original metastore data. Is it possible to do this if I haven't put my metastore data outside the default database (I found information that by default metastore data is stored in derby database on the head node)? Or better way is to create some external database to keep metastore data and point the cluster to use this when re-creating?
Created 04-16-2020 06:35 AM
@Damian_S Yes, I always use mysql/mariadb for hive metastore. If you have the original data you can you just move it to the new location? This should be part of your migration steps regardless of the backend for the metastore.