[ 
https://issues.apache.org/jira/browse/AMQ-6986?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Davide Del Grande updated AMQ-6986:
-----------------------------------
    Description: 
With KahaDB, empty queues (0 pending messages) "survive" broker restarts.

Instead when using JDBC persistence, they just disappear (while queues with 
pending messages work just fine).

 It also seems that Topics are deleted (not sure about this).

 My cfg is:

{{<persistenceAdapter>}}
{{  <jdbcPersistenceAdapter dataDirectory="activemq-data" 
dataSource="#postgres-ds" lockKeepAlivePeriod="5000">}}
{{    <locker>}}
{{      <lease-database-locker leaseHolderId="i-03f2b8bfc2ff67a6d" 
lockAcquireSleepInterval="10000"/>}}
{{    </locker>}}
{{  </jdbcPersistenceAdapter>}}

{{</persistenceAdapter>}}

{{<bean id="postgres-ds" class="org.postgresql.ds.PGPoolingDataSource">}}
{{  <property name="serverName" value="localhost"/>}}
{{  <property name="databaseName" value="activemq"/>}}
{{  <property name="portNumber" value="0"/>}}
{{  <property name="user" value="activemq"/>}}
{{  <property name="password" value="activemq"/>}}
{{  <property name="dataSourceName" value="postgres"/>}}
{{  <property name="initialConnections" value="1"/>}}
{{<property name="maxConnections" value="10"/>}}
{{</bean>}}

 

This is pretty annoying because, at least in my use-case, a monitoring system 
periodically polls queues to see if they are empty or not - and it fails if it 
cannot find the queues that it was polling before broker restart.

And for topics, this can break the application, too.

 

I can't understand if this is per-design (is it documented?) or I'm facing some 
bugs.

 

 

  was:
With KahaDB, empty queues (0 pending messages) "survive" broker restarts.

Instead when using JDBC persistence, they just disappear (while queues with 
pending messages work just fine).

 

It also seems that Topics are deleted (not sure about this).

 

This is pretty annoying because, at least in my use-case, a monitoring system 
periodically polls queues to see if they are empty or not - and it fails if it 
cannot find the queues that it was polling before broker restart.

And for topics, this can break the application, too.

 

I can't understand if this is per-design (is it documented?) or I'm facing some 
bugs.

 

 


> Queues/Topics with no pending messages are lost when using 
> JDBCPersistenceAdapter
> ---------------------------------------------------------------------------------
>
>                 Key: AMQ-6986
>                 URL: https://issues.apache.org/jira/browse/AMQ-6986
>             Project: ActiveMQ
>          Issue Type: Bug
>          Components: JDBC
>    Affects Versions: 5.15.3
>            Reporter: Davide Del Grande
>            Priority: Major
>
> With KahaDB, empty queues (0 pending messages) "survive" broker restarts.
> Instead when using JDBC persistence, they just disappear (while queues with 
> pending messages work just fine).
>  It also seems that Topics are deleted (not sure about this).
>  My cfg is:
> {{<persistenceAdapter>}}
> {{  <jdbcPersistenceAdapter dataDirectory="activemq-data" 
> dataSource="#postgres-ds" lockKeepAlivePeriod="5000">}}
> {{    <locker>}}
> {{      <lease-database-locker leaseHolderId="i-03f2b8bfc2ff67a6d" 
> lockAcquireSleepInterval="10000"/>}}
> {{    </locker>}}
> {{  </jdbcPersistenceAdapter>}}
> {{</persistenceAdapter>}}
> {{<bean id="postgres-ds" class="org.postgresql.ds.PGPoolingDataSource">}}
> {{  <property name="serverName" value="localhost"/>}}
> {{  <property name="databaseName" value="activemq"/>}}
> {{  <property name="portNumber" value="0"/>}}
> {{  <property name="user" value="activemq"/>}}
> {{  <property name="password" value="activemq"/>}}
> {{  <property name="dataSourceName" value="postgres"/>}}
> {{  <property name="initialConnections" value="1"/>}}
> {{<property name="maxConnections" value="10"/>}}
> {{</bean>}}
>  
> This is pretty annoying because, at least in my use-case, a monitoring system 
> periodically polls queues to see if they are empty or not - and it fails if 
> it cannot find the queues that it was polling before broker restart.
> And for topics, this can break the application, too.
>  
> I can't understand if this is per-design (is it documented?) or I'm facing 
> some bugs.
>  
>  



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Reply via email to