lanyuanxiaoyao/hudi
1
0
Fork 0
Code Activity

[HUDI-137] Hudi cleaning state changes should be consistent with compaction actions

Browse Source 
Before this change, Cleaner performs cleaning of old file versions and then stores the deleted files in .clean files.
With this setup, we will not be able to track file deletions if a cleaner fails after deleting files but before writing .clean metadata.
This is fine for regular file-system view generation but Incremental timeline syncing relies on clean/commit/compaction metadata to keep a consistent file-system view.

Cleaner state transitions is now similar to that of compaction.

1. Requested : HoodieWriteClient.scheduleClean() selects the list of files that needs to be deleted and stores them in metadata
2. Inflight : HoodieWriteClient marks the state to be inflight before it starts deleting
3. Completed : HoodieWriteClient marks the state after completing the deletion according to the cleaner plan
This commit is contained in:
Balaji Varadarajan
2019-10-28 18:54:48 -07:00
committed by Balaji Varadarajan
parent 23b303e4b1
commit 1032fc3e54
34 changed files with 856 additions and 491 deletions
Download Patch File Download Diff File Expand all files Collapse all files

41
hudi-hive/src/main/java/org/apache/hudi/hive/HiveSyncTool.java
View File

@@ -62,25 +62,30 @@ public class HiveSyncTool {
}
public void syncHoodieTable() throws ClassNotFoundException {
switch (hoodieHiveClient.getTableType()) {
case COPY_ON_WRITE:
syncHoodieTable(false);
break;
case MERGE_ON_READ:
// sync a RO table for MOR
syncHoodieTable(false);
String originalTableName = cfg.tableName;
// TODO : Make realtime table registration optional using a config param
cfg.tableName = cfg.tableName + SUFFIX_REALTIME_TABLE;
// sync a RT table for MOR
syncHoodieTable(true);
cfg.tableName = originalTableName;
break;
default:
LOG.error("Unknown table type " + hoodieHiveClient.getTableType());
throw new InvalidDatasetException(hoodieHiveClient.getBasePath());
try {
switch (hoodieHiveClient.getTableType()) {
case COPY_ON_WRITE:
syncHoodieTable(false);
break;
case MERGE_ON_READ:
// sync a RO table for MOR
syncHoodieTable(false);
String originalTableName = cfg.tableName;
// TODO : Make realtime table registration optional using a config param
cfg.tableName = cfg.tableName + SUFFIX_REALTIME_TABLE;
// sync a RT table for MOR
syncHoodieTable(true);
cfg.tableName = originalTableName;
break;
default:
LOG.error("Unknown table type " + hoodieHiveClient.getTableType());
throw new InvalidDatasetException(hoodieHiveClient.getBasePath());
}
} catch (RuntimeException re) {
LOG.error("Got runtime exception when hive syncing", re);
} finally {
hoodieHiveClient.close();
}
hoodieHiveClient.close();
}
private void syncHoodieTable(boolean isRealTime) throws ClassNotFoundException {