[
https://issues.apache.org/jira/browse/LUCENE-6110?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
micky.zf updated LUCENE-6110:
-----------------------------
Description:
when the project start running (lt's a WebServer), First will build a
IndexWriter object as single instance.
then every got a request, the method will parse request,then build a
IndexReader .the code seems like this.
....
indexReader = DirectoryReader.open(indexWriter, false);
...
indexReader.close();
return Data;
the engine running, lt's okay. a fews days later, the Performance turnning
down. use jstack shell command . got logs ,
......
"qtp463777123-1058" prio=10 tid=0x00007f716c1a2800 nid=0xb1e1 waiting for
monitor entry [0x00007f6ec0d35000]
java.lang.Thread.State: BLOCKED (on object monitor)
at org.apache.lucene.index.IndexWriter.getReader(IndexWriter.java:411)
- waiting to lock <0x0000000752035af8> (a java.lang.Object)
at
org.apache.lucene.index.DirectoryReader.open(DirectoryReader.java:112)
at
lt's just BLOCKED.
DirectoryReader getReader(boolean applyAllDeletes) throws IOException
ensureOpen();
poolReaders = true;
DirectoryReader r = null;
doBeforeFlush();
boolean anySegmentFlushed = false;
boolean success2 = false;
try {
synchronized (fullFlushLock) { // this line code will be blocked!!!!
.....
was:
when the project start running (lt's a WebServer), First will build a
IndexWriter object as single instance.
then every got a request, the method will parse request,then build a
IndexReader .the code seems like this.
....
indexReader = DirectoryReader.open(indexWriter, false);
...
indexReader.close();
return Data;
the engine running, lt's okay. a fews days later, the Performance turnning
down. use jstack shell command . got logs ,
......
"qtp463777123-1058" prio=10 tid=0x00007f716c1a2800 nid=0xb1e1 waiting for
monitor entry [0x00007f6ec0d35000]
java.lang.Thread.State: BLOCKED (on object monitor)
at org.apache.lucene.index.IndexWriter.getReader(IndexWriter.java:411)
- waiting to lock <0x0000000752035af8> (a java.lang.Object)
at
org.apache.lucene.index.DirectoryReader.open(DirectoryReader.java:112)
at
lt's just BLOCKED.
> When call ' IndexWriter.getReader(IndexWriter) ' BLOCKED..
> -----------------------------------------------------------
>
> Key: LUCENE-6110
> URL: https://issues.apache.org/jira/browse/LUCENE-6110
> Project: Lucene - Core
> Issue Type: Bug
> Components: core/codecs
> Affects Versions: 4.10
> Environment: Description: CentOS release 6.5 (Final)
> Release: 6.5
> Codename: Final
> Reporter: micky.zf
> Labels: performance
> Original Estimate: 96h
> Remaining Estimate: 96h
>
> when the project start running (lt's a WebServer), First will build a
> IndexWriter object as single instance.
> then every got a request, the method will parse request,then build a
> IndexReader .the code seems like this.
> ....
> indexReader = DirectoryReader.open(indexWriter, false);
> ...
> indexReader.close();
> return Data;
> the engine running, lt's okay. a fews days later, the Performance turnning
> down. use jstack shell command . got logs ,
> ......
> "qtp463777123-1058" prio=10 tid=0x00007f716c1a2800 nid=0xb1e1 waiting for
> monitor entry [0x00007f6ec0d35000]
> java.lang.Thread.State: BLOCKED (on object monitor)
> at org.apache.lucene.index.IndexWriter.getReader(IndexWriter.java:411)
> - waiting to lock <0x0000000752035af8> (a java.lang.Object)
> at
> org.apache.lucene.index.DirectoryReader.open(DirectoryReader.java:112)
> at
> lt's just BLOCKED.
> DirectoryReader getReader(boolean applyAllDeletes) throws IOException
> ensureOpen();
> poolReaders = true;
> DirectoryReader r = null;
> doBeforeFlush();
> boolean anySegmentFlushed = false;
> boolean success2 = false;
> try {
> synchronized (fullFlushLock) { // this line code will be blocked!!!!
> .....
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
