rmdmattingly commented on code in PR #5654:
URL: https://github.com/apache/hbase/pull/5654#discussion_r1475316978
##########
hbase-server/src/main/java/org/apache/hadoop/hbase/quotas/DefaultOperationQuota.java:
##########
@@ -132,16 +146,28 @@ public void addMutation(final Mutation mutation) {
operationSize[OperationType.MUTATE.ordinal()] +=
QuotaUtil.calculateMutationSize(mutation);
}
+ @Override
+ public void addBlockBytesScanned(long blockBytesScanned) {
+ this.blockBytesScanned += blockBytesScanned;
+ }
+
/**
* Update estimate quota(read/write size/capacityUnits) which will be
consumed
* @param numWrites the number of write requests
* @param numReads the number of read requests
* @param numScans the number of scan requests
*/
protected void updateEstimateConsumeQuota(int numWrites, int numReads, int
numScans) {
- writeConsumed = estimateConsume(OperationType.MUTATE, numWrites, 100);
- readConsumed = estimateConsume(OperationType.GET, numReads, 100);
- readConsumed += estimateConsume(OperationType.SCAN, numScans, 1000);
+ if (useBlockBytesScanned) {
+ writeConsumed = estimateConsume(OperationType.MUTATE, numWrites, 100);
+ // assume 1 block required for reads. this is probably a low estimate,
which is okay
+ readConsumed = numReads > 0 ? blockSizeBytes : 0;
+ readConsumed += numScans > 0 ? blockSizeBytes : 0;
Review Comment:
After the testing mentioned in
https://github.com/apache/hbase/pull/5654#issue-2101355167, I feel more
confident that these estimates are adequate. We verified that both multiget and
filtered scan workloads get throttled as one would expect.
--
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]
