lanyuanxiaoyao/hudi
1
0
Fork 0
Code Activity

[HUDI-4145] Archives the metadata file in HoodieInstant.State sequence (part2) (#5676)

Browse Source
This commit is contained in:
Danny Chan
2022-05-26 11:21:39 +08:00
committed by GitHub
parent cf837b4900
commit 4e42ed5eae
1 changed files with 31 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files

48
hudi-client/hudi-client-common/src/main/java/org/apache/hudi/client/HoodieTimelineArchiver.java
View File

@@ -506,13 +506,7 @@ public class HoodieTimelineArchiver<T extends HoodieAvroPayload, I, K, O> {
List<HoodieInstant> instantsToStream = groupByTsAction.get(Pair.of(hoodieInstant.getTimestamp(),
HoodieInstant.getComparableAction(hoodieInstant.getAction())));
if (instantsToStream != null) {
// sorts the instants in natural order to make sure the metadata files be removed
// in HoodieInstant.State sequence: requested -> inflight -> completed,
// this is important because when a COMPLETED metadata file is removed first,
// other monitors on the timeline(such as the compaction or clustering services) would
// mistakenly recognize the pending file as a pending operation,
// then all kinds of weird bugs occur.
return instantsToStream.stream().sorted();
return instantsToStream.stream();
} else {
// if a concurrent writer archived the instant
return Stream.empty();
@@ -522,18 +516,28 @@ public class HoodieTimelineArchiver<T extends HoodieAvroPayload, I, K, O> {
private boolean deleteArchivedInstants(List<HoodieInstant> archivedInstants, HoodieEngineContext context) throws IOException {
LOG.info("Deleting instants " + archivedInstants);
boolean success = true;
List<String> instantFiles = archivedInstants.stream().map(archivedInstant ->
new Path(metaClient.getMetaPath(), archivedInstant.getFileName())
).map(Path::toString).collect(Collectors.toList());
List<String> pendingInstantFiles = new ArrayList<>();
List<String> completedInstantFiles = new ArrayList<>();
for (HoodieInstant instant : archivedInstants) {
String filePath = new Path(metaClient.getMetaPath(), instant.getFileName()).toString();
if (instant.isCompleted()) {
completedInstantFiles.add(filePath);
} else {
pendingInstantFiles.add(filePath);
}
}
context.setJobStatus(this.getClass().getSimpleName(), "Delete archived instants: " + config.getTableName());
Map<String, Boolean> resultDeleteInstantFiles = deleteFilesParallelize(metaClient, instantFiles, context, false);
for (Map.Entry<String, Boolean> result : resultDeleteInstantFiles.entrySet()) {
LOG.info("Archived and deleted instant file " + result.getKey() + " : " + result.getValue());
success &= result.getValue();
}
// Delete the metadata files
// in HoodieInstant.State sequence: requested -> inflight -> completed,
// this is important because when a COMPLETED metadata file is removed first,
// other monitors on the timeline(such as the compaction or clustering services) would
// mistakenly recognize the pending file as a pending operation,
// then all kinds of weird bugs occur.
boolean success = deleteArchivedInstantFiles(context, true, pendingInstantFiles);
success &= deleteArchivedInstantFiles(context, success, completedInstantFiles);
// Remove older meta-data from auxiliary path too
Option<HoodieInstant> latestCommitted = Option.fromJavaOptional(archivedInstants.stream().filter(i -> i.isCompleted() && (i.getAction().equals(HoodieTimeline.COMMIT_ACTION)
@@ -545,6 +549,16 @@ public class HoodieTimelineArchiver<T extends HoodieAvroPayload, I, K, O> {
return success;
}
private boolean deleteArchivedInstantFiles(HoodieEngineContext context, boolean success, List<String> files) {
Map<String, Boolean> resultDeleteInstantFiles = deleteFilesParallelize(metaClient, files, context, false);
for (Map.Entry<String, Boolean> result : resultDeleteInstantFiles.entrySet()) {
LOG.info("Archived and deleted instant file " + result.getKey() + " : " + result.getValue());
success &= result.getValue();
}
return success;
}
/**
* Remove older instants from auxiliary meta folder.
*