KnightChess commented on code in PR #9640:
URL: https://github.com/apache/hudi/pull/9640#discussion_r1332664773
##########
hudi-sync/hudi-hive-sync/src/test/java/org/apache/hudi/hive/testutils/HiveTestUtil.java:
##########
@@ -154,6 +155,7 @@ public static void setUp() throws IOException,
InterruptedException, HiveExcepti
hiveSyncProps.setProperty(META_SYNC_PARTITION_FIELDS.key(), "datestr");
hiveSyncProps.setProperty(META_SYNC_PARTITION_EXTRACTOR_CLASS.key(),
SlashEncodedDayPartitionValueExtractor.class.getName());
hiveSyncProps.setProperty(HIVE_BATCH_SYNC_PARTITION_NUM.key(), "3");
+ hiveSyncProps.setProperty(HIVE_AUTO_CREATE_DATABASE.key(), "true");
Review Comment:
yes, usually, there will be no problems with the default creat database,
because user usually will create database first. I will close this pr if it is
not need to change it.
and in out case, we use `waggle_dance` to router multi metastore instances,
and there is something wrong in `waggle_dance` which can not correct connect
the target ms instance when client check db is exist.
for example:
we have two ms instances, is `A` and `B`
and db name `test` create in ms `A`, so every op in db `test` will all
routing to this ms instance `A`.
but something wrong in `waggle_dance` happend, db check use the `B`
instance, and will create a new db `test`, the db is not unique.
may be it has bug in `waggle_dance`, and I think `create db` seem to be
unreasonable in each etl task
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
