s1monw commented on a change in pull request #1397: LUCENE-9304: Refactor
DWPTPool to pool DWPT directly
URL: https://github.com/apache/lucene-solr/pull/1397#discussion_r404068194
##########
File path:
lucene/core/src/java/org/apache/lucene/index/DocumentsWriterFlushControl.java
##########
@@ -380,52 +368,35 @@ DocumentsWriterPerThread nextPendingFlush() {
fullFlush = this.fullFlush;
numPending = this.numPending;
}
- if (numPending > 0 && !fullFlush) { // don't check if we are doing a full
flush
- final int limit = perThreadPool.getActiveThreadStateCount();
- for (int i = 0; i < limit && numPending > 0; i++) {
- final ThreadState next = perThreadPool.getThreadState(i);
- if (next.flushPending) {
- final DocumentsWriterPerThread dwpt = tryCheckoutForFlush(next);
- if (dwpt != null) {
- return dwpt;
+ if (numPending > 0 && fullFlush == false) { // don't check if we are doing
a full flush
+ for (final DocumentsWriterPerThread next : perThreadPool) {
+ if (next.isFlushPending()) {
+ if (next.tryLock()) {
+ try {
+ if (perThreadPool.isRegistered(next)) {
+ return checkOutForFlush(next);
+ }
+ } finally {
+ next.unlock();
Review comment:
so far it's intentional that we are checking out the DWPT of the pool and
unlock it. The reason is that others might block on it to check it out as well
but then won't succeed since it's not part of the pool anymore. I can take
another iteration here and make sure it's locked until flushed. I wonder if we
should do this after the fact.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org
For additional commands, e-mail: issues-h...@lucene.apache.org
