[
https://issues.apache.org/jira/browse/LUCENE-7822?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17098298#comment-17098298
]
Adrien Grand commented on LUCENE-7822:
--------------------------------------
I'd like to revive this issue since I agree with Martin that the
CorruptIndexException is much more useful, the true problem is that the file is
corrupt, not that the version bits are illegal or that a given codec doesn't
exist? In the case of systems like Solr or Elasticsearch, knowing that the
problem is a corruption is also more useful as in some cases you can re-create
the shard from another copy that isn't corrupt automatically.
Would it be acceptable to change checkFooter to throw the CorruptIndexException
with the prior exception as a suppressed exception when checksums don't match?
Or maybe we should stop using the variant of checkFooter that takes a prior
exception and call checksumEntireFile up-front instead on our metadata files as
suggested above?
> IllegalArgumentException thrown instead of a CorruptIndexException
> ------------------------------------------------------------------
>
> Key: LUCENE-7822
> URL: https://issues.apache.org/jira/browse/LUCENE-7822
> Project: Lucene - Core
> Issue Type: Bug
> Affects Versions: 6.5.1
> Reporter: Martin Amirault
> Priority: Minor
> Attachments: LUCENE-7822.patch
>
>
> Similarly to LUCENE-7592 , When an {{*.si}} file is corrupted on very
> specific part an IllegalArgumentException is thrown instead of a
> CorruptIndexException.
> StackTrace (Lucene 6.5.1):
> {code}
> java.lang.IllegalArgumentException: Illegal minor version: 12517381
> at
> __randomizedtesting.SeedInfo.seed([1FEB5987CFA44BE:B8755B5574F9F3BF]:0)
> at org.apache.lucene.util.Version.<init>(Version.java:385)
> at org.apache.lucene.util.Version.<init>(Version.java:371)
> at org.apache.lucene.util.Version.fromBits(Version.java:353)
> at
> org.apache.lucene.codecs.lucene62.Lucene62SegmentInfoFormat.read(Lucene62SegmentInfoFormat.java:97)
> at
> org.apache.lucene.index.SegmentInfos.readCommit(SegmentInfos.java:357)
> at
> org.apache.lucene.index.SegmentInfos.readCommit(SegmentInfos.java:288)
> at org.apache.lucene.index.SegmentInfos$1.doBody(SegmentInfos.java:448)
> at org.apache.lucene.index.SegmentInfos$1.doBody(SegmentInfos.java:445)
> at
> org.apache.lucene.index.SegmentInfos$FindSegmentsFile.run(SegmentInfos.java:692)
> at
> org.apache.lucene.index.SegmentInfos$FindSegmentsFile.run(SegmentInfos.java:644)
> at
> org.apache.lucene.index.SegmentInfos.readLatestCommit(SegmentInfos.java:450)
> at
> org.apache.lucene.index.DirectoryReader.listCommits(DirectoryReader.java:260)
> {code}
> Simple fix would be to add IllegalArgumentException to the catch list at
> {{org/apache/lucene/index/SegmentInfos.java:289}}
> Other variations for the stacktraces:
> {code}
> java.lang.IllegalArgumentException: invalid codec filename '_�.cfs', must
> match: _[a-z0-9]+(_.*)?\..*
> at
> __randomizedtesting.SeedInfo.seed([8B3FDE317B8D634A:A8EE07E5EB4B0B13]:0)
> at
> org.apache.lucene.index.SegmentInfo.checkFileNames(SegmentInfo.java:270)
> at org.apache.lucene.index.SegmentInfo.addFiles(SegmentInfo.java:252)
> at org.apache.lucene.index.SegmentInfo.setFiles(SegmentInfo.java:246)
> at
> org.apache.lucene.codecs.lucene62.Lucene62SegmentInfoFormat.read(Lucene62SegmentInfoFormat.java:248)
> at
> org.apache.lucene.index.SegmentInfos.readCommit(SegmentInfos.java:357)
> at
> org.apache.lucene.index.SegmentInfos.readCommit(SegmentInfos.java:288)
> at org.apache.lucene.index.SegmentInfos$1.doBody(SegmentInfos.java:448)
> at org.apache.lucene.index.SegmentInfos$1.doBody(SegmentInfos.java:445)
> at
> org.apache.lucene.index.SegmentInfos$FindSegmentsFile.run(SegmentInfos.java:692)
> at
> org.apache.lucene.index.SegmentInfos$FindSegmentsFile.run(SegmentInfos.java:644)
> at
> org.apache.lucene.index.SegmentInfos.readLatestCommit(SegmentInfos.java:450)
> at
> org.apache.lucene.index.DirectoryReader.listCommits(DirectoryReader.java:260)
> {code}
> {code}
> java.lang.IllegalArgumentException: An SPI class of type
> org.apache.lucene.codecs.Codec with name 'LucenI62' does not exist. You need
> to add the corresponding JAR file supporting this SPI to your classpath. The
> current classpath supports the following names: [Lucene62, Lucene50,
> Lucene53, Lucene54, Lucene60]
> at
> __randomizedtesting.SeedInfo.seed([925DE160F7260F99:B026EB9373CB6368]:0)
> at org.apache.lucene.util.NamedSPILoader.lookup(NamedSPILoader.java:116)
> at org.apache.lucene.codecs.Codec.forName(Codec.java:116)
> at org.apache.lucene.index.SegmentInfos.readCodec(SegmentInfos.java:424)
> at
> org.apache.lucene.index.SegmentInfos.readCommit(SegmentInfos.java:356)
> at
> org.apache.lucene.index.SegmentInfos.readCommit(SegmentInfos.java:288)
> at org.apache.lucene.index.SegmentInfos$1.doBody(SegmentInfos.java:448)
> at org.apache.lucene.index.SegmentInfos$1.doBody(SegmentInfos.java:445)
> at
> org.apache.lucene.index.SegmentInfos$FindSegmentsFile.run(SegmentInfos.java:692)
> at
> org.apache.lucene.index.SegmentInfos$FindSegmentsFile.run(SegmentInfos.java:644)
> at
> org.apache.lucene.index.SegmentInfos.readLatestCommit(SegmentInfos.java:450)
> at
> org.apache.lucene.index.DirectoryReader.listCommits(DirectoryReader.java:260)
> {code}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org
For additional commands, e-mail: issues-h...@lucene.apache.org
