[jira] Commented: (LUCENE-2324) Per thread DocumentsWriters that write their own private segments

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

[jira] Commented: (LUCENE-2324) Per thread DocumentsWriters that write their own private segments

JIRA jira@apache.org

    [ https://issues.apache.org/jira/browse/LUCENE-2324?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12891664#action_12891664 ]

Michael Busch commented on LUCENE-2324:
---------------------------------------

{quote}
Well, though, if we did write it "live", since the doc is marked deleted, it would be "harmless" right?

Only difference is it took up some disk space (which should be minor given that it's not "normal" to have lots of docs hitting non-aborting exceptions).

Also... when the doc is large, the fact that we double buffer (write first to RAMFile then to the store) can be costly.  Ie we consume 1X the stored-field size in transient RAM, vs writing directly to disk, I think?
{quote]

Yeah I totally agree that this would be a nice performance win - I got first excited too about removing that extra buffering layer.  But we probably have to handle exceptions a bit differently then?  Because now all exceptions thrown in processDocument() are considered non-aborting, because they nothing is actually written to disk in processDocument().

We only abort when we encounter an exception in finishDocument(), because then the files might be corrupted.

So we probably just have to catch the IOExceptions somewhere else to figure out if it was e.g. a (non-aborting) exception from the TokenStream vs. e.g. a disk full (aborting) exception from StoredFieldsWriter.

> Per thread DocumentsWriters that write their own private segments
> -----------------------------------------------------------------
>
>                 Key: LUCENE-2324
>                 URL: https://issues.apache.org/jira/browse/LUCENE-2324
>             Project: Lucene - Java
>          Issue Type: Improvement
>          Components: Index
>            Reporter: Michael Busch
>            Assignee: Michael Busch
>            Priority: Minor
>             Fix For: Realtime Branch
>
>         Attachments: lucene-2324.patch, lucene-2324.patch, LUCENE-2324.patch
>
>
> See LUCENE-2293 for motivation and more details.
> I'm copying here Mike's summary he posted on 2293:
> Change the approach for how we buffer in RAM to a more isolated
> approach, whereby IW has N fully independent RAM segments
> in-process and when a doc needs to be indexed it's added to one of
> them. Each segment would also write its own doc stores and
> "normal" segment merging (not the inefficient merge we now do on
> flush) would merge them. This should be a good simplification in
> the chain (eg maybe we can remove the *PerThread classes). The
> segments can flush independently, letting us make much better
> concurrent use of IO & CPU.

--
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: [hidden email]
For additional commands, e-mail: [hidden email]