[
https://issues.apache.org/jira/browse/LUCENE-2471?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12967205#action_12967205
]
Robert Muir commented on LUCENE-2471:
-------------------------------------
I think the problem actually wasn't interrupting but some sort of race
condition?
Either way, I think its sad we had to disable the optimization, and it would be
nice if we put it back (safely). But some notes:
* an NIO-based optimization shouldnt be in FSDirectory, instead in
NIOFSDirectory! I don't think SimpleFS should use NIO methods.
* the real problem is our tests. we unfortunately don't have many tests that
search with multiple threads. This is why it was so hard to reproduce the bug.
* we should make sure we benchmark any change to ensure its faster because
there is unfortunately a lot of risk in optimizing this method.
> Supporting bulk copies in Directory
> -----------------------------------
>
> Key: LUCENE-2471
> URL: https://issues.apache.org/jira/browse/LUCENE-2471
> Project: Lucene - Java
> Issue Type: Improvement
> Components: Store
> Reporter: Earwin Burrfoot
> Fix For: 3.1, 4.0
>
>
> A method can be added to IndexOutput that accepts IndexInput, and writes
> bytes using it as a source.
> This should be used for bulk-merge cases (offhand - norms, docstores?). Some
> Directories can then override default impl and skip intermediate buffers
> (NIO, MMap, RAM?).
--
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]
