[
https://issues.apache.org/jira/browse/LUCENE-2324?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12979138#action_12979138
]
Jason Rutherglen commented on LUCENE-2324:
------------------------------------------
{quote}So all that's guaranteed after the global flush() returns is that all
state present prior to when flush() is invoked, is moved to disk. Ie if addDocs
are still happening concurrently then the DWPTs will start filling up again
even while the "global flush" runs. That's fine.{quote}
What if the user wants a guaranteed hard flush of all state up to the point of
the flush call (won't they want this sometimes with getReader)? If we're
flushing sequentially (without pausing all threads) we're removing that? Maybe
we'll need to give the option of global lock/stop or sequential flush?
Also I think we need to clear the thread bindings of a DWPT just prior to the
flush of the DWPT? Otherwise (when multiple threads are mapped to a single
DWPT) the other threads will wait on the [main] DWPT flush when they should be
spinning up a new DWPT?
Then, what happens to reusing the DWPT if we're flushing it, and we spin a new
DWPT (effectively replacing the old DWPT), eg, we're going to lose the byte[]
recycling? Maybe we need to and share and sync the byte[] pooling between DWPTs
or will that noticeably affect indexing performance?
> 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-SMALL.patch, LUCENE-2324-SMALL.patch,
> LUCENE-2324-SMALL.patch, LUCENE-2324-SMALL.patch, LUCENE-2324-SMALL.patch,
> lucene-2324.patch, lucene-2324.patch, LUCENE-2324.patch, test.out, test.out
>
>
> 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: [email protected]
For additional commands, e-mail: [email protected]
