Hi,
For Kylin 5, you have to use a distributed storage, and the default
choice is HDFS,
and alternative choice is Cloud Storage(like S3), you can NOT deploy and
run Kylin
without a distributed storage.
Besides, you need a RDBMS as a metastore, Zookeeper as service
discovery,
a Spark cluster as compute service, a Hive Metastore for seeking
databases and tables.
Finally, HBae is totally NOT necessary for Kylin 4.0 or higher.
For the question 'Could I use Kylin with just MySQL + Sqoop? (no Hive)
', the
answer is no, you need to install and deployed Zookeeper, a distributed
storage
(HDFS or cloud storage), a Spark cluster and a Hive metastore. Here is a
diagram
maybe helpful:
https://kylin.apache.org/images/blog/kylin4_on_cloud/3_kylin_cluster.jpg
Here are some links:
- https://kylin.apache.org/blog/2022/04/20/kylin4-on-cloud-part1/
-
https://kylin.apache.org/5.0/docs/deployment/on-premises/installation/platform/install_on_apache_hadoop
------------------------
With warm regard
Xiaoxiang Yu
On Sat, Aug 26, 2023 at 8:03 PM marc nicole <[email protected]> wrote:
> Hello,
>
> I have few questions regarding storage mean for Kylin:
>
> I was wondering if Kylin would work normally if I don't configure it to
> work with any storage tool (as MySQL with Sqoop Or with Hive)? It would
> then automatically use HDFS ?
>
> Also is configuring HBASE necessary?
>
> Could I use Kylin with just MySQL + Sqoop? (no Hive)
> What the use of HBase if the normal used storage is Hive?
>
> Thanks. Regards
>
