Fix disk usage estimation for SORTED_SET doc values.

martijnvg · martijnvg · commit 6b81b37df22b · 2025-08-28T19:35:32.000+02:00
The `nextOrd()` was never invoked, which can hugely underestimate bytes used for doc sorted set doc values.
diff --git a/server/src/main/java/org/elasticsearch/action/admin/indices/diskusage/IndexDiskUsageAnalyzer.java b/server/src/main/java/org/elasticsearch/action/admin/indices/diskusage/IndexDiskUsageAnalyzer.java
@@ -277,6 +277,7 @@ void analyzeDocValues(SegmentReader reader, IndexDiskUsageStats stats) throws IO
                     SortedSetDocValues sortedSet = iterateDocValues(maxDocs, () -> docValuesReader.getSortedSet(field), dv -> {
                         for (int i = 0; i < dv.docValueCount(); i++) {
                             cancellationChecker.logEvent();
+                            dv.nextOrd();
                         }
                     });
                     if (sortedSet.getValueCount() > 0) {

Original file line number	Diff line number	Diff line change
`@@ -277,6 +277,7 @@ void analyzeDocValues(SegmentReader reader, IndexDiskUsageStats stats) throws IO`
`277`	`277`	`SortedSetDocValues sortedSet = iterateDocValues(maxDocs, () -> docValuesReader.getSortedSet(field), dv -> {`
`278`	`278`	`for (int i = 0; i < dv.docValueCount(); i++) {`
`279`	`279`	`cancellationChecker.logEvent();`
	`280`	`+ dv.nextOrd();`
`280`	`281`	`}`
`281`	`282`	`});`
`282`	`283`	`if (sortedSet.getValueCount() > 0) {`