[
https://issues.apache.org/jira/browse/LUCENE-6341?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14350155#comment-14350155
]
Michael McCandless commented on LUCENE-6341:
--------------------------------------------
This is awesome.
+1 for patch and 5.x.
Does -verbose work with -fast? I think it should (we seem to do null checks
for all the terms dict stats), maybe add a test? It's a nice (fast0 way to see
RAM usage for the index...
Does -exorcise and -fast work?
In the usage can you also confess that the identifiers are also "cross-checked"?
> add CheckIndex -fast option
> ---------------------------
>
> Key: LUCENE-6341
> URL: https://issues.apache.org/jira/browse/LUCENE-6341
> Project: Lucene - Core
> Issue Type: Improvement
> Reporter: Robert Muir
> Attachments: LUCENE-6341.patch
>
>
> CheckIndex is great for testing and when tracking down lucene bugs.
> But in cases where users just want to verify their index files are OK, it is
> very slow and expensive.
> I think we should add a -fast option, that only opens the reader and calls
> checkIntegrity(). This means all files are the correct files (identifiers
> match) and have the correct CRC32 checksums.
> For our 10M doc wikipedia index, this is the difference between a 2 second
> check and a 2 minute check.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
