danny0405 commented on code in PR #10874:
URL: https://github.com/apache/hudi/pull/10874#discussion_r1527740554
##########
hudi-client/hudi-client-common/src/main/java/org/apache/hudi/metadata/HoodieBackedTableMetadataWriter.java:
##########
@@ -1407,43 +1382,7 @@ protected void cleanIfNecessary(BaseHoodieWriteClient
writeClient, String instan
writeClient.lazyRollbackFailedIndexing();
}
- /**
- * Validates the timeline for both main and metadata tables to ensure
compaction on MDT can be scheduled.
- */
- protected boolean validateTimelineBeforeSchedulingCompaction(Option<String>
inFlightInstantTimestamp, String latestDeltaCommitTimeInMetadataTable) {
- // we need to find if there are any inflights in data table timeline
before or equal to the latest delta commit in metadata table.
- // Whenever you want to change this logic, please ensure all below
scenarios are considered.
- // a. There could be a chance that latest delta commit in MDT is committed
in MDT, but failed in DT. And so findInstantsBeforeOrEquals() should be employed
- // b. There could be DT inflights after latest delta commit in MDT and we
are ok with it. bcoz, the contract is, the latest compaction instant time in
MDT represents
- // any instants before that is already synced with metadata table.
- // c. Do consider out of order commits. For eg, c4 from DT could complete
before c3. and we can't trigger compaction in MDT with c4 as base instant time,
until every
- // instant before c4 is synced with metadata table.
- List<HoodieInstant> pendingInstants =
dataMetaClient.reloadActiveTimeline().filterInflightsAndRequested()
-
.findInstantsBeforeOrEquals(latestDeltaCommitTimeInMetadataTable).getInstants();
-
- if (!pendingInstants.isEmpty()) {
- checkNumDeltaCommits(metadataMetaClient,
dataWriteConfig.getMetadataConfig().getMaxNumDeltacommitsWhenPending());
- LOG.info(String.format(
- "Cannot compact metadata table as there are %d inflight instants in
data table before latest deltacommit in metadata table: %s. Inflight instants
in data table: %s",
- pendingInstants.size(), latestDeltaCommitTimeInMetadataTable,
Arrays.toString(pendingInstants.toArray())));
- return false;
- }
-
- // Check if there are any pending compaction or log compaction instants in
the timeline.
- // If pending compact/logCompaction operations are found abort scheduling
new compaction/logCompaction operations.
- Option<HoodieInstant> pendingLogCompactionInstant =
Review Comment:
Hi, @suryaprasanna , can you help confirm this part, in
`HoodieLogCompactionPlanGenerator` we already skip scheduling compaction plan
for any pending compaction/logcompaction file slices, so why we still need to
cancel the scheduling the compaction here?
--
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]
