[kulbhushanc]

Yes it was a clean install and main server group didnt  report this, as mentioned only server three having this issue

[shawkins]

The other server group uses the full-ha profile.  The domain script provided by Teiid only installs into the ha profile.  Do you also expect Teiid to be in the full-ha profile, and/or are you seeing the admin client automatically assume that you are using the full-ha profile?

[kulbhushanc]

Yes Steven... I know teiid only be working with ha profile, It is hard-coded in the code itself, that is why I changed profile of other-server-group to ha.

Now Server groups in my domain.xml :-

<server-groups>

        <server-group name="main-server-group" profile="full">

            <jvm name="default">

                <heap size="64m" max-size="512m"/>

            </jvm>

            <socket-binding-group ref="full-sockets"/>

        </server-group>

        <server-group name="other-server-group" profile="ha">

            <jvm name="default">

                <heap size="64m" max-size="512m"/>

            </jvm>

            <socket-binding-group ref="ha-sockets"/>

        </server-group>

    </server-groups>

After changing <socket-binding-group ref="full-ha-sockets"/> to <socket-binding-group ref="ha-sockets"/> form other-server-group I am not getting any error during the start-up or load time of the server.

but code line

connection = DriverManager.getConnection("jdbc:teiid:" + VDBName + "@mm://" + IPAddress + ":" + port + ";", userName, password);

still throws an error:

org.teiid.jdbc.TeiidSQLException: TEIID20020 Error establishing socket to host and port: <IP_Address>:31000. Reason: Connection refused: connect

[shawkins]

> still throws an error:

This is introducing several new changes.  Let's focus on having a common understanding on the other issues you have seen first.

> Yes Steven... I know teiid only be working with ha profile, It is hard-coded in the code itself,

So to clarify are you expecting Teiid to install into both the ha and full-ha?

And how are you connecting the AdminApi - are you using the AdminFactory.createAdmin call that allows you to specify the profile name, or connecting with the AdminShell such that it's assuming which profile to use?

[kulbhushanc]

I am creating object of the class AdminApiClientAccessor by calling AdminApiClientAccessor.getInstance() method which interns returns

new AdminApiClientAccessor()

Code:

private AdminApiClientAccessor clientAccessor = AdminApiClientAccessor.getInstance();

then using clientAccessor I am able to access VDB(s) and their information like is it active or not etc..

I am creating Admin using code line

admin = AdminFactory.getInstance().createAdmin(ModelControllerClient.Factory.create(serverHost, serverPort, cbh));

[shawkins]

There is also a createAdmin method that takes a profileName parameter.  You can control which profile you use with that.  Otherwise you'll default to ha.

[kulbhushanc]

Thanks Steven....

How should I deal with the error?

org.teiid.jdbc.TeiidSQLException: TEIID20020 Error establishing socket to host and port: <IP_Address>:31000. Reason: Connection refused: connect

[shawkins]

Can you outline exactly the steps you did to alter the configuration?  And did you look at the server3 log to confirm what port it's using - perhaps a port offset is being applied.

[kulbhushanc]

Can you outline exactly the steps you did to alter the configuration >

AS7 Cluster Howto - JBoss AS 7.1 - Project Documentation Editor

https://docs.jboss.org/author/display/SWITCHYARD/Domain+Clustering+with+SwitchYard

I have mentioned in domain.xml file under the tag

<socket-binding-group name="ha-sockets" default-interface="public">

.....

<socket-binding name="teiid-jdbc" port="31000"/>

<socket-binding name="teiid-odbc" port="35432"/>

</socket-binding-group>

port-offset for server3 is 250

did you look at the server3 log to confirm what port it's using >

Yes it is 31250, for the teiid instance on master when I access teiid using this port number I am able to make the connection to the database; as I am using mod_cluster I need to access teiid port irrespective of the teiid instance I mean whether its a master or slave, but now when I shut down the master I am not able to access the teiid because slave teiid instance is running on a different port.

[shawkins]

> but now when I shut down the master I am not able to access the teiid because slave teiid instance is running on a different port.

So to make sure I understand correctly you are saying that this scenario works as expected while the master node is running, but does not work when the master EAP node is shutdown and httpd/mod_cluster is still running?

[kulbhushanc]

Yes, right...

[shawkins]

What do you see in the mod_cluster-manager status page?

[kulbhushanc]

When there is both node connected mod_cluster-manager status page looks like:

mod_cluster/1.3.1.Final

Auto Refresh show DUMP output show INFO output

Node slave_local:server-three-slave (ajp://<IP_Addr>:8459):

Enable Contexts Disable Contexts Stop Contexts

Balancer: main-server-group,LBGroup: ,Flushpackets: Off,Flushwait: 10000,Ping: 10000000,Smax: 65,Ttl: 60000000,Status: OK,Elected: 0,Read: 0,Transferred: 0,Connected: 0,Load: 100

Virtual Host 1:

Contexts:

/zclusterDemo, Status: ENABLED Request: 0 Disable Stop

Aliases:

localhost

default-host

Node master:server-three (ajp://<IP_Addr>:8259):

Enable Contexts Disable Contexts Stop Contexts

Balancer: main-server-group,LBGroup: ,Flushpackets: Off,Flushwait: 10000,Ping: 10000000,Smax: 65,Ttl: 60000000,Status: OK,Elected: 0,Read: 0,Transferred: 0,Connected: 0,Load: 100

Virtual Host 1:

Contexts:

/zclusterDemo, Status: ENABLED Request: 0 Disable Stop

Aliases:

localhost

default-host

where zclusterDemo is the war file I deployed using admin api. it can be accessible from both the node i.e. when I shutdown the master or vice versa I can still access the zclusterDemo application.

[shawkins]

I can't say what may be the issue at first glance.  It's probably a good idea to start with a fresh post in the WildFly forum as well.  I'll have to try to set this up locally to reproduce.