sanjeet006py commented on code in PR #7136:
URL: https://github.com/apache/hbase/pull/7136#discussion_r2185574409
##########
hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/RegionScannerImpl.java:
##########
@@ -145,7 +150,18 @@ private static boolean hasNonce(HRegion region, long
nonce) {
} finally {
region.smallestReadPointCalcLock.unlock(ReadPointCalculationLock.LockType.RECORDING_LOCK);
}
+ boolean isScanMetricsEnabled = scan.isScanMetricsEnabled();
+
ThreadLocalServerSideScanMetrics.setScanMetricsEnabled(isScanMetricsEnabled);
+ if (isScanMetricsEnabled) {
+ ThreadLocalServerSideScanMetrics.reset();
+ }
initializeScanners(scan, additionalScanners);
+ if (isScanMetricsEnabled) {
+ bytesReadFromFs +=
ThreadLocalServerSideScanMetrics.getBytesReadFromFsAndReset();
Review Comment:
Reads happening during StoreScanner init are not special in any way in my
opinion. They also add to the latency like how any other block seek will add.
So, I don't see a value add in tracking them separately. Also, for tracking
bytesRead for trailer and load-on-open section, as that will happen outside of
user triggered read path so, there also I feel we don't need to track
separately. WDYT?
--
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]
