Huang Kuan Hao created HDDS-16216:
-------------------------------------
Summary: Avoid the flatten-copy in
KeyInputStream.getBlockLocationInfo
Key: HDDS-16216
URL: https://issues.apache.org/jira/browse/HDDS-16216
Project: Apache Ozone
Issue Type: Sub-task
Reporter: Huang Kuan Hao
Assignee: Huang Kuan Hao
getBlockLocationInfo allocates a flattened list on the per-block read /
pipeline-refresh path just to find one block by ID:
List<OmKeyLocationInfo> collect =
newKeyInfo.getLatestVersionLocations()
.getLocationList()
.stream()
.filter(l -> l.getBlockID().equals(blockID))
.collect(Collectors.toList());
if (CollectionUtils.isNotEmpty(collect)) {
return collect.get(0);
} else {
return null;
}
Fix: iterate getLocationLists() and return on the first blockID match
(short-circuits, no copy). Do not use getBlocksLatestVersionOnly() — the block
may live in a non-latest createVersion bucket, so narrowing the search would
change behavior. Behavior unchanged.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
