3 Replies Latest reply on Aug 27, 2014 11:25 AM by Laurent Henssen

    Upgrade from 6.2 to 6.3 : mandatory to upgrade both DC and HC at the same time ?

    Laurent Henssen Newbie

      Hello everybody,

       

       

      I just have a little question.

       

      We have 1 DC for several HC (for different environment).

       

      Previously, I upgraded DC from 6.1 to 6.2 and HC could stay at 6.1 or upgrade too to 6.2.

       

      With EAP 6.3, HC is failing.

       

       

      Here is an example of failing for HC 6.2 to DC 6.3 (both are fresh and no customized configuration has been implemented yet : only the minimum => authentication) :

       

      =========================================================================
      
      
        JBoss Bootstrap Environment
      
      
        JBOSS_HOME: /home/laurent/jboss/6.3_dev/jboss-eap-6.2
      
      
        JAVA: java
      
      
        JAVA_OPTS: -Xms64m -Xmx512m -XX:MaxPermSize=256m -Djava.net.preferIPv4Stack=true -Djboss.modules.system.pkgs=org.jboss.byteman -Djava.awt.headless=true
      
      
      =========================================================================
      
      
      12:42:58,994 INFO  [org.jboss.modules] (main) JBoss Modules version 1.3.0.Final-redhat-2
      12:42:59,138 INFO  [org.jboss.as.process.Host Controller.status] (main) JBAS012017: Starting process 'Host Controller'
      [Host Controller] 12:42:59,747 INFO  [org.jboss.modules] (main) JBoss Modules version 1.3.0.Final-redhat-2
      [Host Controller] 12:42:59,998 INFO  [org.jboss.msc] (main) JBoss MSC version 1.0.4.GA-redhat-1
      [Host Controller] 12:43:00,092 INFO  [org.jboss.as] (MSC service thread 1-4) JBAS015899: JBoss EAP 6.2.0.GA (AS 7.3.0.Final-redhat-14) starting
      [Host Controller] 12:43:00,923 INFO  [org.xnio] (MSC service thread 1-3) XNIO Version 3.0.7.GA-redhat-1
      [Host Controller] 12:43:00,946 INFO  [org.xnio.nio] (MSC service thread 1-3) XNIO NIO Implementation Version 3.0.7.GA-redhat-1
      [Host Controller] 12:43:00,972 INFO  [org.jboss.remoting] (MSC service thread 1-3) JBoss Remoting version 3.2.18.GA-redhat-1
      [Host Controller] 12:43:01,056 INFO  [org.jboss.as.remoting] (MSC service thread 1-2) JBAS017100: Listening on 127.0.0.1:9999
      [Host Controller] 12:43:02,769 ERROR [org.jboss.as.host.controller] (Controller Boot Thread) JBAS010901: Could not connect to master. Aborting. Error was: java.io.IOException: 1-$-JBAS014893: Transforming resource [
      [Host Controller]     ("profile" => "default"),
      [Host Controller]     ("subsystem" => "logging"),
      [Host Controller]     ("periodic-rotating-file-handler" => "FILE")
      [Host Controller] ] for host controller 'wk01lhe' to subsystem 'logging' model version '1.3.0' --there were problems with some of the attributes and this resource will need to be ignored on that host. Details of problems: [JBAS014895: The following attributes are not understood in the target model version and this resource will need to be ignored on the target host: [named-formatter]]
      [Host Controller] 12:43:02,797 INFO  [org.jboss.as] (MSC service thread 1-2) JBAS015950: JBoss EAP 6.2.0.GA (AS 7.3.0.Final-redhat-14) stopped in 20ms
      12:43:03,118 INFO  [org.jboss.as.process.Host Controller.status] (reaper for Host Controller) JBAS012010: Process 'Host Controller' finished with an exit status of 99
      12:43:03,119 INFO  [org.jboss.as.process] (Thread-8) JBAS012016: Shutting down process controller
      12:43:03,120 INFO  [org.jboss.as.process] (Thread-8) JBAS012015: All processes finished; exiting
      
      

       

       

      So, is that a "normal" behavior ? Must I upgrade DC and HC at the same time ? What if you have several environment (not acceptable to upgrade all environment at the same time => nonsense) ?

       

       

      Thanks

        • 1. Re: Upgrade from 6.2 to 6.3 : mandatory to upgrade both DC and HC at the same time ?
          Wolf-Dieter Fink Master

          The problem is that you need to have a compatible configuration.

           

          It is recommended to update the DC first, if so you need to keep the 'old' configuration to not add new features.

          After the DC you can update all the HC controllers.

          At the end you can change the configuration and add new features or fixes with the management interface.

          • 2. Re: Re: Upgrade from 6.2 to 6.3 : mandatory to upgrade both DC and HC at the same time ?
            Laurent Henssen Newbie

            Hello Wolf,

             

             

            > It is recommended to update the DC first, if so you need to keep the 'old' configuration to not add new features.

            That's exactly what I've done. The above error come from a fresh test with no new configuration at all.

            Or yes, may be, the DC started with it's new provided default configuration which could cause an issue => my test is not correct.

             

            I'm currently speaking with RH to solve that issue. The fact is that we have 1 DC and 3 HC running EAP 6.2.4.

            When I tried to update the DC (this time, with configuration from DC 6.2.4), it had about the same issue.

             

            I just re-do that test for RH :

            - unzip new DC 6.3.0

            - copy domain/ from 6.2.4 to DC 6.3.0

            - start DC (complains about messaging subsystem => destination queue must now be added, if not, WARN :

             

            [Host Controller] 13:22:25,239 WARN  [org.jboss.messaging] (Controller Boot Thread) JBAS011618: There is no resource matching the expiry-address jms.queue.ExpiryQueue for the address-settings #, expired messages from destinations matching this address-setting will be lost!
            [Host Controller] 13:22:25,240 WARN  [org.jboss.messaging] (Controller Boot Thread) JBAS011619: There is no resource matching the dead-letter-address jms.queue.DLQ for the address-settings #, undelivered messages from destinations matching this address-setting will be lost!
            

             

            => not an issue, DC is up and running 6.3.0

             

            - copy HC 6.2.4 (from currently running and connected to DC 6.2.4)

            - clean up following folders : log/ domain/log domain/data domain/servers domain/tmp

            - start HC and connect it to new DC 6.3.0 : failure :

             

            [Host Controller] 13:32:26,633 ERROR [org.jboss.as.host.controller] (Controller Boot Thread) JBAS010901: Could not connect to master. Aborting. Error was: java.io.IOException: 1-$-JBAS014893: Transforming resource [
            [Host Controller]    ("profile" => "full_dev"),
            [Host Controller]    ("subsystem" => "datasources"),
            [Host Controller]    ("data-source" => "ldm-logback")
            [Host Controller] ] for host controller 'srv-jboss6dev01' to subsystem 'datasources' model version '1.2.0' --there were problems with some of the attributes and this resource will need to be ignored on that host. Details of problems: [JBAS010486: Attributes [statistics-enabled] rejected. Must be true]
            [Host Controller] 13:32:26,684 INFO  [org.jboss.as] (MSC service thread 1-3) JBAS015950: JBoss EAP 6.2.4.GA (AS 7.3.4.Final-redhat-1) stopped in 46ms
            13:32:27,006 INFO  [org.jboss.as.process.Host Controller.status] (reaper for Host Controller) JBAS012010: Process 'Host Controller' finished with an exit status of 99
            13:32:27,007 INFO  [org.jboss.as.process] (Thread-8) JBAS012016: Shutting down process controller
            13:32:27,008 INFO  [org.jboss.as.process] (Thread-8) JBAS012015: All processes finished; exiting
            

             

            So, old configuration from 6.2.4 was well pushed to new DC 6.3.0 => no new feature at all.

            HC leave with error : I really don't know what I must do now.

             

             

            So, any idea ?

            Last time, I tried to remove datasources ... and then, it was another subsystem which was failing ...

            Fix that datasource should be a good start ;-)

             

             

            The failing DS :

             

                                <datasource jndi-name="java:/ldm-logback" pool-name="ldm-logback" enabled="true" use-ccm="false">
                                    <connection-url>jdbc:sqlserver://${ldm-logback_db_mssql_hostname_dev}:${ldm-logback_db_mssql_port_dev};DatabaseName=${ldm-logback_db_mssql_database_dev};applicationName=ldm-logback</connection-url>
                                    <driver>MSSQL_4-4.0</driver>
                                    <pool>
                                        <min-pool-size>0</min-pool-size>
                                        <max-pool-size>80</max-pool-size>
                                    </pool>
                                    <security>
                                        <user-name>${ldm-logback_db_mssql_username_dev}</user-name>
                                        <password>${ldm-logback_db_mssql_password_dev}</password>
                                    </security>
                                    <validation>
                                        <valid-connection-checker class-name="org.jboss.jca.adapters.jdbc.extensions.mssql.MSSQLValidConnectionChecker"/>
                                        <check-valid-connection-sql>SELECT 1</check-valid-connection-sql>
                                        <validate-on-match>false</validate-on-match>
                                        <background-validation>false</background-validation>
                                    </validation>
                                    <timeout>
                                        <idle-timeout-minutes>15</idle-timeout-minutes>
                                        <allocation-retry>5</allocation-retry>
                                        <allocation-retry-wait-millis>1000</allocation-retry-wait-millis>
                                    </timeout>
                                    <statement>
                                        <share-prepared-statements>false</share-prepared-statements>
                                    </statement>
                                </datasource>
            

             

            Please note that I added 60 new datasources few time ago. They are now yet in use, but HC doesn't complain on startup (version 6.2.4).

             

            Thanks

            • 3. Re: Re: Re: Upgrade from 6.2 to 6.3 : mandatory to upgrade both DC and HC at the same time ?
              Laurent Henssen Newbie

              Hi,

               

               

              We found the issue ...

               

              There is a property called "statistics-enabled".

              On 6.2, it is set to "true" internaly.

              On 6.3, it is now set to "false" internaly.

               

              The new DC, running 6.3, push a "false" to HC 6.2 => it fails.

              Simply set that attribute (on datasource) to false like this (temporarily, just the time to get all HC of the DC upgraded) :

                                  <datasource jndi-name="java:/ldm-logback" pool-name="ldm-logback" enabled="true" use-ccm="false" statistics-enabled="true">
              

               

               

              Now, HC can connect (have to fix all datasources).

               

              Thanks to RH.

               

              Solved!