我在像这样的ANT脚本中从CloverETL调用runGraph类
<java classname="org.jetel.main.runGraph" fork="true" failonerror="true">
<arg value="${canadapostdir}/metadata/CanadaPostADDTypes34.grf"/>
<classpath>
<pathelement path="${ear.lib}/clover.jar" />
<pathelement path="${ear.libext}/oracle.jdbc/ojdbc6.jar" />
<pathelement path="${ear.libext}/mysql.jdbc/mysql-connector-java-3.0.15-ga-bin.jar" />
<fileset dir="${ear.lib}">
<include name="*.jar"/>
</fileset>
</classpath>
</java>
当三叶草开始时,我可以在日志中看到:
[java] [Clover] starting WatchDog thread ...
[java] [WatchDog] Thread started.
[java] [WatchDog] Running on 16 CPU(s) max available memory for JVM 3005703 kB
[java] [Clover] Initializing phase: 0
[java] all edges initialized successfully...
[java] initializing nodes:
[java] INPUT ...OK
[java] TYPEFILTER ...OK
[java] REF ...OK
[java] OUTPUT ...OK
[java] [Clover] phase: 0 initialized successfully.
[java] [WatchDog] Starting up all nodes in phase [0]
[java] [WatchDog] INPUT ... started
[java] [WatchDog] TYPEFILTER ... started
[java] [WatchDog] REF ... started
[java] [WatchDog] OUTPUT ... started
[java] [WatchDog] Sucessfully started all nodes in phase!
我的问题是:如何使用比3005703 kB更多的内存来运行三叶草?服务器上实际上有100gig的免费ram正在运行此脚本。
编辑:我知道我看到三叶草没有在一个单独的jvm上运行,并且它使用了当前JVM实例上可用的所有可用内存。我需要找到一种方法来在一个单独的jvm上启动三叶草,我可以通过-Xms10240m -Xmx10240m。答案 0 :(得分:3)
<jvmarg>
嵌套元素用于传递JVM参数:
<java classname="org.jetel.main.runGraph" fork="true" failonerror="true">
<arg value="${canadapostdir}/metadata/CanadaPostADDTypes34.grf"/>
<jvmarg value="-Xms1024M"/>
<jvmarg value="-Xmx10240M"/>
<classpath>
<pathelement path="${ear.lib}/clover.jar" />
<pathelement path="${ear.libext}/oracle.jdbc/ojdbc6.jar" />
<pathelement path="${ear.libext}/mysql.jdbc/mysql-connector-java-3.0.15-ga-bin.jar" />
<fileset dir="${ear.lib}">
<include name="*.jar"/>
</fileset>
</classpath>
</java>
您正在在另一个VM中执行该类。这就是fork=true
的含义。