keith-turner commented on code in PR #4761:
URL: https://github.com/apache/accumulo/pull/4761#discussion_r1693270304
##########
core/src/main/java/org/apache/accumulo/core/iterators/SynchronizedServerFilter.java:
##########
@@ -48,13 +49,19 @@ public
SynchronizedServerFilter(SortedKeyValueIterator<Key,Value> source) {
@Override
public synchronized void next() throws IOException {
+ if (Thread.interrupted()) {
Review Comment:
> Looking at NativeMap and SortedMapIterator, it appears the pattern is to
check the interrupt on every seek and every 100th next.
What about checking the thread interrupt status in those places and in
RFile? That same places where the boolean is checked, those places are meant
to run less frequently. A long long time ago I did performance test and could
see differences in the number of key values per second read depending on how
frequently the volatile was checked (suspected it was interfering w/ the
processor cache, but was not sure). The check of the atomic boolean in RFile
is done when a new data block is read so that is naturally less frequent.
NativeMap code does the periodic check thing you mentioned.
--
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]
