[
https://issues.apache.org/activemq/browse/AMQ-1380?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=41168#action_41168
]
palmsun edited comment on AMQ-1380 at 1/27/08 9:54 PM:
-------------------------------------------------------------
This patch is still not part of ActiveMQ 5.0, where slave file handle leaks
are still an issue. I manually applied this patch and it fixed the problem. Why
is this patch not part of 5.0? Running a master/slave (persistence based on
journal file-system) using a conf file below, leads to the crash of the slave
after a few hours, due to running out of open files. If you monitor the number
of open files (lsof | wc -l), you'll see that the number of open files keeps
increasing every 10 seconds. With this patch, this is no-longer a problem.
Please include this patch in the release. Here is the conf file that I've used
for master/slave (the only difference is that the value of the IP address in
the broker name and the transport connector)
<beans
xmlns="http://www.springframework.org/schema/beans";
xmlns:amq="http://activemq.org/config/1.0";
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance";
xsi:schemaLocation="http://www.springframework.org/schema/beans
http://www.springframework.org/schema/beans/spring-beans-2.0.xsd
http://activemq.org/config/1.0
http://activemq.apache.org/schema/activemq-core.xsd
http://activemq.apache.org/camel/schema/spring
http://activemq.apache.org/camel/schema/spring/camel-spring.xsd";>
<!-- Allows us to use system properties as variables in this configuration
file -->
<bean
class="org.springframework.beans.factory.config.PropertyPlaceholderConfigurer"/>
<broker xmlns="http://activemq.org/config/1.0"; brokerName="10.0.1.200"
useJmx="false">
<!-- The transport connectors ActiveMQ will listen to -->
<transportConnectors>
<transportConnector name="openwire"
uri="tcp://10.0.1.200:61616?trace=true"/>
</transportConnectors>
<persistenceAdapter>
<journaledJDBC dataDirectory="/Users/sjavey/amq-broker-data"/>
</persistenceAdapter>
</broker>
<!-- An embedded servlet engine for serving up the Admin console -->
<jetty xmlns="http://mortbay.com/schemas/jetty/1.0";>
<connectors>
<nioConnector port="8161" />
</connectors>
<handlers>
<webAppContext contextPath="/admin"
resourceBase="${activemq.base}/webapps/admin" logUrlOnStart="true" />
</handlers>
</jetty>
</beans>
was (Author: palmsun):
This patch is still not part of ActiveMQ 5.0, where slave file handle
leaks are still an issue. I manually applied this patch and it fixed the
problem. Why is this patch not part of 5.0? Running a master/slave (persistence
based on journal file-system) using a conf file below, leads to the crash of
the slave after a few hours, due to running out of open files. If you monitor
the number of open files (lsof | wc -l), you'll see that the number of open
files keeps increasing every 10 seconds. With this patch, this is no-longer a
problem. Please include this patch in the release.
<beans
xmlns="http://www.springframework.org/schema/beans";
xmlns:amq="http://activemq.org/config/1.0";
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance";
xsi:schemaLocation="http://www.springframework.org/schema/beans
http://www.springframework.org/schema/beans/spring-beans-2.0.xsd
http://activemq.org/config/1.0
http://activemq.apache.org/schema/activemq-core.xsd
http://activemq.apache.org/camel/schema/spring
http://activemq.apache.org/camel/schema/spring/camel-spring.xsd";>
<!-- Allows us to use system properties as variables in this configuration
file -->
<bean
class="org.springframework.beans.factory.config.PropertyPlaceholderConfigurer"/>
<broker xmlns="http://activemq.org/config/1.0"; brokerName="10.0.1.200"
useJmx="false">
<!-- The transport connectors ActiveMQ will listen to -->
<transportConnectors>
<transportConnector name="openwire"
uri="tcp://10.0.1.200:61616?trace=true"/>
</transportConnectors>
<persistenceAdapter>
<journaledJDBC dataDirectory="/Users/sjavey/amq-broker-data"/>
</persistenceAdapter>
</broker>
<!-- An embedded servlet engine for serving up the Admin console -->
<jetty xmlns="http://mortbay.com/schemas/jetty/1.0";>
<connectors>
<nioConnector port="8161" />
</connectors>
<handlers>
<webAppContext contextPath="/admin"
resourceBase="${activemq.base}/webapps/admin" logUrlOnStart="true" />
</handlers>
</jetty>
</beans>
> ActiveIO does not close ControlFile
> ------------------------------------
>
> Key: AMQ-1380
> URL: https://issues.apache.org/activemq/browse/AMQ-1380
> Project: ActiveMQ
> Issue Type: Bug
> Components: Broker
> Affects Versions: 4.1.1
> Environment: All
> Reporter: Joe Kutner
> Attachments: patch.txt
>
>
> The following exception is thrown from an ActiveMQ instance that is in Slave
> mode that causes it to crash.
> java.io.FileNotFoundException: ***/*******/*****/*******/journal/control.dat
> (Too many open files)
> at java.io.RandomAccessFile.open(Native Method)
> at java.io.RandomAccessFile.<init>(RandomAccessFile.java:212)
> at
> org.apache.activeio.journal.active.ControlFile.<init>(ControlFile.java:60)
> at
> org.apache.activeio.journal.active.LogFileManager.initialize(LogFileManager.java:119)
> at
> org.apache.activeio.journal.active.LogFileManager.<init>(LogFileManager.java:102)
> at
> org.apache.activeio.journal.active.JournalImpl.<init>(JournalImpl.java:101)
> at
> org.apache.activemq.store.DefaultPersistenceAdapterFactory.createJournal(DefaultPersistenceAdapterFactory.java:224)
> at
> org.apache.activemq.store.DefaultPersistenceAdapterFactory.getJournal(DefaultPersistenceAdapterFactory.java:132)
> at
> org.apache.activemq.store.DefaultPersistenceAdapterFactory.createPersistenceAdapter(DefaultPersistenceAdapterFactory.java:70)
> at
> org.apache.activemq.store.PersistenceAdapterFactoryBean.getObject(PersistenceAdapterFactoryBean.java:35)
> at
> org.springframework.beans.factory.support.AbstractBeanFactory.getObjectFromFactoryBean(AbstractBeanFactory.java:994)
> ... 41 more
> The problem is that each time ActiveIO checks to see if control.dat is locked
> it creates a new java.io.RandomAccessFile. If the control.dat file is
> locked, it fails to close the RandomAccessFile.
> This exception always will be thrown from an instance in Shared-File-System
> Slave mode that continues to check the Journal file (which the Master has
> locked). But the rate at which is is thrown depends on the maximum number of
> files the OS allows to be open. You can check how many files are open with
> 'ulimit -a'
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
