codope commented on a change in pull request #4693:
URL: https://github.com/apache/hudi/pull/4693#discussion_r838136779
##########
File path:
hudi-client/hudi-client-common/src/main/java/org/apache/hudi/metadata/HoodieBackedTableMetadataWriter.java
##########
@@ -511,24 +523,42 @@ private boolean
initializeFromFilesystem(HoodieTableMetaClient dataMetaClient,
initializeMetaClient(dataWriteConfig.getMetadataConfig().populateMetaFields());
initTableMetadata();
- initializeEnabledFileGroups(dataMetaClient, createInstantTime);
+ // if async metadata indexing is enabled,
+ // then only initialize files partition as other partitions will be built
using HoodieIndexer
+ List<MetadataPartitionType> enabledPartitionTypes = new ArrayList<>();
+ if (dataWriteConfig.isMetadataAsyncIndex()) {
Review comment:
I did keep it that way before i.e. you could build files partition as
well using the indexer. However, there was a suggestion earlier that files
partition should always be indexed inline
https://github.com/apache/hudi/pull/4693#discussion_r824509601 (due to its
critical nature).
So i pivoted to keeping files inline and only allow the other partitions
async.
Another secondary point was that files partitions takes much less time,
other partitions could take a lot of time to build for large tables, so keep
them disabled by default on the inline path.
imo, from user standpoint, files partition can also be indexed async.
--
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]
