> On Nov 3, 2017, at 12:08 PM, Stack <[email protected]> wrote:
>
> On Sat, Oct 28, 2017 at 2:00 PM, Konstantin Shvachko <[email protected]>
> wrote:
>
>> It is an interesting question whether Ozone should be a part of Hadoop.
>
> I don't see a direct answer to this question. Is there one? Pardon me if
> I've not seen it but I'm interested in the response.
+1
Given:
* a completely different set of config files (ozone-site.xml, etc)
* package name is org.apache.hadoop.ozone, not
org.apache.hadoop.hdfs.ozone
… it doesn’t really seem to want to be part of HDFS, much less Hadoop.
Plus hadoop-hdfs-project/hadoop-hdfs is already a battle zone when it comes to
unit tests, dependencies, etc [*]
At a minimum, it should at least be using it’s own maven module for a
lot of the bits that generates it’s own maven jars so that we can split this
functionality up at build/test time.
At a higher level, this feels a lot like the design decisions that were
made around yarn-native-services. This feature is either part of HDFS or it’s
not. Pick one. Doing both is incredibly confusing for everyone outside of the
branch.
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
