An index copied over another index can result in corruption
-----------------------------------------------------------
Key: LUCENE-1691
URL: https://issues.apache.org/jira/browse/LUCENE-1691
Project: Lucene - Java
Issue Type: Improvement
Components: Store
Reporter: Adrian Hempel
Priority: Minor
Fix For: 2.4.1
After restoring an older backup of an index over the top of a newer version of
the index, attempts to open the index can result in CorruptIndexExceptions,
such as:
{noformat}
Caused by: org.apache.lucene.index.CorruptIndexException: doc counts differ for
segment _ed: fieldsReader shows 1137 but segmentInfo shows 1389
at org.apache.lucene.index.SegmentReader.initialize(SegmentReader.java:362)
at org.apache.lucene.index.SegmentReader.get(SegmentReader.java:306)
at org.apache.lucene.index.SegmentReader.get(SegmentReader.java:228)
at
org.apache.lucene.index.MultiSegmentReader.<init>(MultiSegmentReader.java:55)
at
org.apache.lucene.index.ReadOnlyMultiSegmentReader.<init>(ReadOnlyMultiSegmentReader.java:27)
at
org.apache.lucene.index.DirectoryIndexReader$1.doBody(DirectoryIndexReader.java:102)
at
org.apache.lucene.index.SegmentInfos$FindSegmentsFile.run(SegmentInfos.java:653)
at
org.apache.lucene.index.DirectoryIndexReader.open(DirectoryIndexReader.java:115)
at org.apache.lucene.index.IndexReader.open(IndexReader.java:316)
at org.apache.lucene.index.IndexReader.open(IndexReader.java:237)
{noformat}
The apparent cause is the strategy of taking the maximum of the ID in the
segments.gen file, and the IDs of the apparently valid segment files (See lines
523-593
[here|http://svn.apache.org/viewvc/lucene/java/tags/lucene_2_4_1/src/java/org/apache/lucene/index/SegmentInfos.java?annotate=751393]),
and using this as the current generation of the index. This will include
"stale" segments that existed before the backup was restored.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
