Yes, as far as I know, Measure relies on Livy server, Livy server relies on 
Spark which have metastore configurations. it is make sense because Livy server 
is managed by spark administrators.


So it seems to be another option to get hive meta through rest api from Livy 
server. 


On the other hand, I think griffin is the platform for end user to do data 
quality, so the platform should not have any authority to access any data in 
Hadoop. The datasource should be configured by end user so that griffin can 
submit the data quality job.


Best Wish
Jason Liao




------------------ ???????? ------------------
??????: "Nick Sokolov"<[email protected]>;
????????: 2018??12??9??(??????) ????1:28
??????: "dev"<[email protected]>;

????: Re: [DISCUSS] hive server2 vs hive metastore



Is that just for Service or for Measure as well? As far as understand,
Spark relies on direct metastore availability, and there is no good way
around it for Measure.

If that's just for Service, it might be tricky to parse out all column
metadata correctly, as there might be differences in the way how particular
hive versions format output. There seems to be another option -- WebHCat
exposes database and table information in API
<https://cwiki.apache.org/confluence/display/Hive/WebHCat+Reference> and
it's supposed to be available for end users.

In either case it feels like this mechanism should be pluggable, to
maintain compatibility with existing installations.

On Sat, Dec 8, 2018 at 7:14 PM William Guo <[email protected]> wrote:

> Hi ALL,
>
> Normally, the administrator of hive won't to open hive metastore server to
> users, because it is not safety for the data in hive warehouse. So, can we
> use hive server2 instead of hive metastore server to get hive meta.
>
> https://issues.apache.org/jira/browse/GRIFFIN-222
>
> Do you think we should use hive server2 for coming release?
>
> Thanks,
> William
>

Reply via email to