AmatyaAvadhanula commented on code in PR #15952:
URL: https://github.com/apache/druid/pull/15952#discussion_r1502197054
##########
server/src/main/java/org/apache/druid/metadata/SegmentsMetadataManagerConfig.java:
##########
@@ -31,6 +31,9 @@ public class SegmentsMetadataManagerConfig
@JsonProperty
private Period pollDuration = new Period("PT1M");
+ @JsonProperty
+ private boolean useFindFullyOvershadowed;
Review Comment:
The new flow is empirically better for large clusters with thousands of
segments per interval.
There may not be much of a benefit for smaller clusters and I had added the
config to allow a rollback to the current flow since the methods in the new
flow are not well documented.
An alternative that I was thinking of was to detect if an interval has
segment-locking enabled i.e if there are segments with distinct minor versions.
If not, we optimize the existing isOvershadowed call to work faster with
timechunk-locking based segments. Does this seem like a better approach?
--
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]
