dsmiley commented on code in PR #2666:
URL: https://github.com/apache/solr/pull/2666#discussion_r1776344004
##########
solr/CHANGES.txt:
##########
@@ -109,8 +109,11 @@ New Features
Improvements
---------------------
+* SOLR-17158: Users for whom partial results are uninteresting may set
partialResultsAllowed=false. This allows Solr to reduce time spent processing
partial results and omit them from the response. (Gus Heck, Andrzej Bialeki,
hossman)
Review Comment:
I don't understand "for whom partial results are uninteresting", especially
that word "uninteresting". Meaning, "for whom partial results aren't
acceptable"?
##########
solr/core/src/java/org/apache/solr/handler/component/ParallelHttpShardHandler.java:
##########
@@ -41,57 +39,78 @@
@NotThreadSafe
public class ParallelHttpShardHandler extends HttpShardHandler {
+ @SuppressWarnings("unused")
private static final Logger log =
LoggerFactory.getLogger(MethodHandles.lookup().lookupClass());
private final ExecutorService commExecutor;
+ /*
+ * Unlike the basic HttpShardHandler, this class allows us to exit submit
before
+ * pending is incremented and the responseFutureMap is updated. If the
runnables that
Review Comment:
Ooof, the additional complexity I'm reading about is concerning. Without
looking at this as closely as either of you, admittedly, I've naively wondered
why doesn't HttpShardHandler use ExecutorCompletionService. It's a really nice
JDK utility that seems to do conceptually what HttpShardHandler is doing. If
it would be helpful, it would reduce the complexity of what this Solr class has
to deal with. Thoughts?
> we need to know that everything we scheduled has been attempted (fail or
not) before we let the loop that takes responses exit or we might miss one
That doesn't sound right. Meaning, I hope we can cancel if we haven't
actually sent all requests.
--
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]
