[
https://issues.apache.org/jira/browse/LUCENE-2061?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Michael McCandless updated LUCENE-2061:
---------------------------------------
Attachment: LUCENE-2061.patch
New nrtBench.py attached, fixed a few small issues... also, I removed
-Xbatch to java; it seems to make less consistent results.
My initial results:
JAVA:
java version "1.6.0_14"
Java(TM) SE Runtime Environment (build 1.6.0_14-b08)
Java HotSpot(TM) 64-Bit Server VM (build 14.0-b16, mixed mode)
OS:
SunOS rhumba 5.11 snv_111b i86pc i386 i86pc Solaris
Baseline QPS 158.12
||Indexing docs/sec||NRT reopen period (sec)||QPS add||QPS update||QPS add (%
diff)||QPS update (% diff)||
|1|1|157.5|125.7|{color:red}-0.4%{color}|{color:red}-20.5%{color}|
|1|2.5|157.6|127.5|{color:red}-0.4%{color}|{color:red}-19.4%{color}|
|1|5|156.9|127.2|{color:red}-0.8%{color}|{color:red}-19.5%{color}|
|10|0.1|156.3|142.4|{color:red}-1.2%{color}|{color:red}-9.9%{color}|
|10|0.5|155.8|125.0|{color:red}-1.5%{color}|{color:red}-20.9%{color}|
|10|1|156.0|142.6|{color:red}-1.3%{color}|{color:red}-9.8%{color}|
|10|2.5|156.6|143.4|{color:red}-0.9%{color}|{color:red}-9.3%{color}|
|10|5|156.2|144.0|{color:red}-1.2%{color}|{color:red}-8.9%{color}|
|100|0.1|153.9|138.8|{color:red}-2.7%{color}|{color:red}-12.2%{color}|
|100|0.5|155.0|141.1|{color:red}-2.0%{color}|{color:red}-10.8%{color}|
|100|1|156.1|141.3|{color:red}-1.3%{color}|{color:red}-10.6%{color}|
|100|2.5|155.9|116.7|{color:red}-1.4%{color}|{color:red}-26.2%{color}|
|100|5|157.0|143.8|{color:red}-0.7%{color}|{color:red}-9.1%{color}|
|1000|0.1|145.9|110.0|{color:red}-7.7%{color}|{color:red}-30.4%{color}|
|1000|0.5|148.0|117.6|{color:red}-6.4%{color}|{color:red}-25.6%{color}|
|1000|1|148.3|97.7|{color:red}-6.2%{color}|{color:red}-38.2%{color}|
|1000|2.5|149.3|99.1|{color:red}-5.6%{color}|{color:red}-37.3%{color}|
|1000|5|147.4|124.3|{color:red}-6.8%{color}|{color:red}-21.4%{color}|
The docs are ~1KB sized docs derived from wikipedia. The searching is
only running a single fixed query (1), over and over.
Some rough observations:
* Even at only 1 update/sec, QPS already drops way too much
(~20%), which is weird. Something is amiss.
* At all indexing rates, handling updates slows NRT down much more
than pure adds.
* Pure adds (no deletes) are handled quite well, at <= 100 adds/sec,
the hit to QPS is ~1-2%, which is great. Even at 1000 docs/sec
the QPS only drops ~6%, which seems reasonable.
* There isn't really a clear correlation of reopen rate to QPS,
which is also weird.
Looks like we have some puzzles to solve...
> Create benchmark & approach for testing Lucene's near real-time performance
> ---------------------------------------------------------------------------
>
> Key: LUCENE-2061
> URL: https://issues.apache.org/jira/browse/LUCENE-2061
> Project: Lucene - Java
> Issue Type: Task
> Components: Index
> Reporter: Michael McCandless
> Assignee: Michael McCandless
> Priority: Minor
> Attachments: LUCENE-2061.patch, LUCENE-2061.patch
>
>
> With the improvements to contrib/benchmark in LUCENE-2050, it's now
> possible to create compelling algs to test indexing & searching
> throughput against a periodically reopened near-real-time reader from
> the IndexWriter.
> Coming out of the discussions in LUCENE-1526, I think to properly
> characterize NRT, we should measure net search throughput as a
> function of both reopen rate (ie how often you get a new NRT reader
> from the writer) and indexing rate. We should also separately measure
> pure adds vs updates (deletes + adds); the latter is much more work
> for Lucene.
> This can help apps make capacity decisions... and can help us test
> performance of pending improvements for NRT (eg LUCENE-1313,
> LUCENE-2047).
--
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]
