lpavanvenkat commented on code in PR #11685:
URL: https://github.com/apache/ozone/pull/11685#discussion_r4583801634
##########
hadoop-ozone/recon/src/main/java/org/apache/hadoop/ozone/recon/tasks/MultipartInfoInsightHandler.java:
##########
@@ -129,6 +130,7 @@ public void handleDeleteEvent(OMDBUpdateEvent<String,
Object> event, String tabl
objectCountMap.computeIfPresent(getTableCountKeyFromTable(tableName),
(k, count) -> count > 0 ? count - 1L : 0L);
applyLegacyPartSizes(multipartKeyInfo, tableName, unReplicatedSizeMap,
replicatedSizeMap, false);
+ cacheSplitSchemaReplicationConfig(multipartKeyInfo);
Review Comment:
multipartPartsTable rows have no replication config, so part sizes use the
parent MPU's config, cached by uploadId. The cache is empty after a Recon
restart. When an upload started before the restart is completed or aborted,
only the multipartInfoTable DELETE arrives before its part DELETEs. Without
caching on DELETE, those part sizes were never subtracted from the MPU totals.
--
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]