Alexander Kolbasov commented on HIVE-18685:

Thank you for proposing this. A few questions about the current design.
 # Should we also consider sharding by catalog so that each catalog could be 
served by different server instance or served by different backing DB? 
Otherwise we'll have common issues of contending over the same DB locks.
 # How would this work with notifications? DO you envision a single stream of 
notifications which just includes catalog in the message or separate streams 
per catalog? Note that notifications are currently  strictly serialized using 
DB lock.
 # Would you consider possibility of each catalog implemented by potentially 
different backing store?
 # Your document doesn't describe any restrictions on catalog name - should it 
be limited by size or allowed character set?
 # Do you envision catalog having some kind of properties (outside of security 
model) that can be shared by all databases within a catalog?
 # Why do you need a separate CTLG_NAME table if catalog names are directly 
stored in DBS tables?
 # You mentioned catalog location - what is it used for?
 # Do you think that catalog name would be somehow reflected in HDFS object 
names or not?
 # Is there a need to explicitly create 'hive' catalog - can catalogs be 
created on demand?
 # Is there a need to change Thrift API to support catalogs - would it be 
possible to add notion of catalog by adding some kind of special character to 
db name (something like catalog#database)?
 # If we start adding new APIs should we use this as a possibility to clean up 
APIs or we'll just multiply current API space by 2?



> Add catalogs to metastore
> -------------------------
>                 Key: HIVE-18685
>                 URL: https://issues.apache.org/jira/browse/HIVE-18685
>             Project: Hive
>          Issue Type: New Feature
>          Components: Metastore
>    Affects Versions: 3.0.0
>            Reporter: Alan Gates
>            Assignee: Alan Gates
>            Priority: Major
>         Attachments: HMS Catalog Design Doc.pdf
> SQL supports two levels of namespaces, called in the spec catalogs and 
> schemas (with schema being equivalent to Hive's database).  I propose to add 
> the upper level of catalog.  The attached design doc covers the use cases, 
> requirements, and brief discussion of how it will be implemented in a 
> backwards compatible way.

This message was sent by Atlassian JIRA

Reply via email to