danny0405 commented on code in PR #11809:
URL: https://github.com/apache/hudi/pull/11809#discussion_r1726746601
##########
hudi-spark-datasource/hudi-spark/src/main/scala/org/apache/spark/sql/hudi/command/procedures/RunClusteringProcedure.scala:
##########
@@ -159,7 +150,12 @@ class RunClusteringProcedure extends BaseProcedure
try {
client = HoodieCLIUtils.createHoodieWriteClient(sparkSession, basePath,
confs,
tableName.asInstanceOf[Option[String]])
-
+ if (metaClient.getTableConfig.isMetadataTableAvailable) {
+ if (!confs.contains(HoodieLockConfig.LOCK_PROVIDER_CLASS_NAME.key)) {
+ confs = confs ++ HoodieCLIUtils.getLockOptions(basePath,
metaClient.getBasePath.toUri.getScheme,
client.getConfig.getCommonConfig.getProps())
+ logInfo("Auto config filesystem lock provider for metadata table")
Review Comment:
So the new logic become this when s3 is used and the lock provider config is
missing: the auto generated lock provider config is empty here, but we still
log a msg to indicate a fs lock provider is configured.
BTW, I check the `StorageSchemes` and only HDFS supports atomic creation of
files, but this seems not correct, HDFS relies on `RENAME` to keep atomicity,
while file creation is already atomic for most of the object storage like S3,
GCS ..
--
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]
