[GitHub] [hudi] zhuanshenbsj1 commented on a diff in pull request #7891: [HUDI-5728] HoodieTimelineArchiver archives the latest instant before inflight replacecommit

["zhuanshenbsj1 (via GitHub)" <gi...@apache.org>]

zhuanshenbsj1 commented on code in PR #7891:
URL: https://github.com/apache/hudi/pull/7891#discussion_r1102382018


##########
hudi-client/hudi-client-common/src/main/java/org/apache/hudi/client/HoodieTimelineArchiver.java:
##########
@@ -473,6 +473,33 @@ private Stream<HoodieInstant> getCommitInstantsToArchive() throws IOException {
                       HoodieTimeline.compareTimestamps(s.getTimestamp(), LESSER_THAN, instantToRetain.getTimestamp()))
                   .orElse(true)
           );
+
+      // When inline or async clustering is enabled, we need to ensure that there is a commit in the active timeline
+      // to check whether the file slice generated in pending clustering after archive isn't committed
+      // via {@code HoodieFileGroup#isFileSliceCommitted(slice)}
+      boolean isOldestPendingReplaceInstant =
+          oldestPendingCompactionAndReplaceInstant.map(instant ->
+              HoodieTimeline.REPLACE_COMMIT_ACTION.equals(instant.getAction())).orElse(false);
+      if (isOldestPendingReplaceInstant) {
+        List<HoodieInstant> instantsToArchive = instantToArchiveStream.collect(Collectors.toList());
+        Option<HoodieInstant> latestInstantRetainForReplace = Option.fromJavaOptional(
+            instantsToArchive.stream()
+                .filter(s -> HoodieTimeline.compareTimestamps(
+                    s.getTimestamp(),
+                    LESSER_THAN,
+                    oldestPendingCompactionAndReplaceInstant.get().getTimestamp()))
+                .reduce((i1, i2) -> i2));
+        if (latestInstantRetainForReplace.isPresent()) {
+          LOG.info(String.format(
+              "Retaining the archived instant %s before the inflight replacecommit %s.",
+              latestInstantRetainForReplace.get().getTimestamp(),
+              oldestPendingCompactionAndReplaceInstant.get().getTimestamp()));
+        }
+        instantToArchiveStream = instantsToArchive.stream()
+            .filter(s -> latestInstantRetainForReplace.map(instant -> s.compareTo(instant) != 0)
+                .orElse(true));
+      }
+

Review Comment:
   How about move these changes to function getOldestInstantToRetainForClustering, It looks a little messy.



-- 
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: commits-unsubscribe@hudi.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org