jasonk000 commented on code in PR #14639:
URL: https://github.com/apache/druid/pull/14639#discussion_r1271324520
##########
server/src/main/java/org/apache/druid/metadata/IndexerSQLMetadataStorageCoordinator.java:
##########
@@ -1760,8 +1761,22 @@ public void updateSegmentMetadata(final Set<DataSegment>
segments)
@Override
public Void inTransaction(Handle handle, TransactionStatus
transactionStatus) throws Exception
{
+ final String updatePayload = StringUtils.format("UPDATE %s SET
payload = :payload WHERE id = :id", dbTables.getSegmentsTable());
+ final PreparedBatch batch = handle.prepareBatch(updatePayload);
+
for (final DataSegment segment : segments) {
Review Comment:
I'm OK with this from a performance standpoint, BUT, it changes the
semantics of the code. In the existing code, the delete and update passes are
fully committed, OR failed and rolled back as a unit. If we switch to
transaction batches, then we lose the rollback of the whole piece of work and
may see partially completed work. But, if partial completion is OK, we might be
better to avoid transactions altogether since there's no atomicity requirement.
Thoughts?
--
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
