Multiple multi-datasource to failover

We have a heavy architecture in mind:
Multiple servers would connect to one Oracle RAC in a datacenter. Another Oracle RAC would be set in Standby in another datacenter for Disaster and Recovery.
If the first datacenter is down, the idea is to automatically promote the RAC of the other datacenter as the Primary, so that the application can switch to connect to the other Oracle RAC.
On Weblogic 10.3.2, I know it is possible to failover between nodes of a multi datasource.
However, is it possible to failover from one multi-datasource (pointing the the first Oracle RAC) to another one (pointing to the second Oracle RAC) when the first multi datasource is down?
We do not want to open connections on the second multi-datasouce if the first one is active.
I hope this is clear...
Thank you

But what if, instead, we defined the first Datasource to connect to all the nodes of the first RAC:
jdbc:oracle:thin:@(DESCRIPTION=
(ADDRESS_LIST=(ADDRESS=(PROTOCOL=TCP)(HOST=*host1*)(PORT=1521))
(ADDRESS=(PROTOCOL=TCP)(HOST=*host2*)(PORT=1521)))
(LOAD_BALANCE=yes)
(CONNECT_DATA=(SERVER=DEDICATED)(SERVICE_NAME=SID)
(FAILOVER_MODE=(TYPE=select)(METHOD=basic)(RETRIES=5)(DELAY=2))))
and the other datasource to connect to all the servers of the second RAC?
jdbc:oracle:thin:@(DESCRIPTION=
(ADDRESS_LIST=(ADDRESS=(PROTOCOL=TCP)(HOST=*host3*)(PORT=1521))
(ADDRESS=(PROTOCOL=TCP)(HOST=*host4*)(PORT=1521)))
(LOAD_BALANCE=yes)
(CONNECT_DATA=(SERVER=DEDICATED)(SERVICE_NAME=SID)
(FAILOVER_MODE=(TYPE=select)(METHOD=basic)(RETRIES=5)(DELAY=2))))
And to have the multi datasource to have the failover protocol managed between the two datasources.
Would it be possible?
Is it what you were mentioning, Ravish Mody?
Thanks,
Pierre
Edited by: user3415297 on Apr 8, 2011 8:27 AM
Edited by: user3415297 on Apr 8, 2011 8:33 AM

Similar Messages

Add gridlink datasource to Multi Datasource for failover in weblogic

We have a RAC installation for which I have configured a Gridlink datasource. Also we have a single node DB which is used as a failover database in case RAC is down.
Now I want to configure a Multi Datasource so that I can add the Gridlink datasource as well as the Generic Datasource in the Multi Datasource and configure it for failover. But when I create the Multi Datasource, I do not see the option to add Gridlink datasource to Multidatasource.
Can you please suggest what can be done to add the gridlink datasource to multidatasource? If this is not possible, then how can the Generic single node data source be configured as a failover datasource in case the Gridlink Datasource is down.
Thanks in advance

What is the problem you are trying to address? Is the issue the availability/non-availability of your RAC database? It may make sense to focus on this first.
The normal use case for Multi Data Source is failover or load balancing between nodes of a highly available database, such as RAC which keeps data synchronised.
I am not sure if using a Multi Data Source across Oracle RAC and a non RAC database is supported. If the intention is to provide a HA solution, you will need to have a data synchronisation solution as well. As this is what Oracle RAC does, we come back to my opening question.........
I assume that Multi Data Sources cannot include Gridlink and non-Gridlink sources for these reasons and also because Gridlink is intended as a replacement for Multi Data Sources, providing better failover, load balancing and performance.
Hope this is helpful
Mark

JDBC Configuration with Multi Datasource

Hi all,
We're using Weblogic 9.2 and has configured Multi Datasource with Failover algorithm. All is working well, but our DBA has reported that some requests were randomly still going to the 2nd node although he was sure that nothing was wrong with the 1st node. Is this an expected behavior with the Failover algorithm.
Any help would be greatly appreciated.
Thanks & Regards,
Besetya

Some of the application-level work? I think you can configure an MDS to 'fail over' if the first pool is
simply at capacity and unable to deliver another connection. You can check that, and make sure the
first pool is big enough. Than, also configure the backup pool with an initial capacity of 0, so it
never even makes a connection unless the MDS fails over to it.

Load balancing with multi datasource

Hi,
I have created two data sources and created a multi data source with those with "Load-Balancing" set as algorithm.
I have two kinds of urls where I can use those links to test the load balancing
1. survey page which will take only 2 sec to submit
2. registration page which will take more than 10 sec to submit.
My problem here is everything is fine with survey page and also sequence is in order in the database but with the registration page, all data is getting inserted through the first data source.
Thanks,

Did you check the test connections in reserve option and also failover when busy option for the datasources and multi datasource?

Multi datasource faliover problem - JDeveloper integrated Weblogic 10.3.5

Hi all,
I have a problem with JDeveloper integrated weblogic (version 10.3.5) fail over to second datasource in multi data source (2 generic datasources) configuration environment.
My test goes like this:
1. I have two databases running on separate servers.
2. both data sources are up & running OK.
3. JDeveloper integrated weblogic configured with multi datasource, using algorithm failover, and failover callback class.
4. weblogic started and adf application running using first datasource in list.
5. if I unplug network cable or shut down first database, weblogic won't fail over to second datasource and still trying to use the first one.
Failover callback class:
package failover;
public class ConnectionPoolFailoverCallback implements weblogic.jdbc.extensions.ConnectionPoolFailoverCallback {
     static {
          System.out.println("------------------------------------------------------------");
          System.out.println("\n\n\tConnectionPoolFailoverCallback class initiated...\n\n\t");
          System.out.println("------------------------------------------------------------");
     public int allowPoolFailover(String currPool, String nextPool, int opcode) {
          System.out.println("Current Pool - " + currPool);
          System.out.println("Next Pool - " + nextPool);
          System.out.println("Operation Code - " + opcode);
          System.out.println("------------------------------------------------------------");
          if (opcode == OPCODE_CURR_POOL_DEAD) {
               System.out.println("\n\n\tConnectionPoolFailoverCallback     return OK.");
               return OK;
          } else {
               System.out.println("\n\n\tConnectionPoolFailoverCallback     return DONOT_FAILOVER.");
               return DONOT_FAILOVER;
}{code}
*Problem:* When I unplug network cable or shutdown first database, method allowPoolFailover returns OK, but weblogic won't switch to second datasource entering allowPoolFailover method again and again.
+Note:+ weblogic integrated in JDeveloper is running u development mode.
Did I miss something in configuration or something else?
Any help is appreciated.
Thanks in advance,
Alexandar

Hi Alexandar,
Have you made any progress with this? We seem to have bumped into this issue as well. Connection requests are consistently routed to the disabled pool (it appears as disabled in the WLS console, but we can see the counters increasing steadily), the callback handler is called and returns OK, and the connection request is ultimately serviced by the active pool. So while the application still successfully handles requests, there's obviously an underlying problem in obtaining the connection.
Thanks,
Sebastien

ORA-01002: fetch out of sequence using multiple XA datasources

Hi,
I have a problem accessing multiple XA datasources :
- launch an sql request on datasource 1
- rs.next() on the resultset
- use the data to launch another sql request on datasource 2
After 10 iterations, the next() method throws an SQLException (ORA-01002: fetch out of sequence).
After further investigation, I noticed that :
- the problem doesn't occur if the same datasource is used for the 2 requests
- if I set the fetch size to 15 for the first request, the exception is thrown after 15 iterations (but I can't use this as a workaround because the number is potentially above the million).
Anyone experiencing the same problem ?
Thanks in advance
Nicolas
Here's my configuration :
- Weblogic 8.1 SP4
- JDK sun 1.4.2_04 (we have the same problem with various JDKs including JRockit)
- Oracle 10g
- Oracle thin driver xa 10.1.0.2.0
A JSP I use to reproduce the problem :
<pre><%@ page language="java" contentType="text/html; charset=ISO-8859-1" pageEncoding="ISO-8859-1"%>
<%@ page import="java.util.ArrayList,
                    java.sql.Connection,
                    java.sql.PreparedStatement,
                    java.sql.ResultSet,
                    java.sql.SQLException,
                    javax.naming.InitialContext,
                    javax.naming.NamingException,
                    javax.sql.DataSource,
                    javax.transaction.UserTransaction"%>
<%!
     public void launchTest() throws Exception {
          Connection cnx =null ;
          try {
               getUserTransaction().begin();
               // SQL access #1
               cnx = getConnection("jdbc/xaDatasource1");
               PreparedStatement stmt = cnx.prepareStatement("SELECT test_col from test_table");
               ResultSet rs = stmt.executeQuery();
               // iterate on resulset from SQL 1
               int i=1;
               while (rs.next()){ // SQL exception after 10 iterations
                    System.out.println(i + "");
                    i++;
                    // SQL access #2
                    Connection cnx2 = getConnection("jdbc/xaDatasource2");
                    // problem occurs even if we don't request
                    cnx2.close();
                    // end SQL access #2
               getUserTransaction().commit();
          } catch (Exception e) {
               e.printStackTrace();
               try {
                    getUserTransaction().rollback();
               } catch (Exception e1) {
                    e1.printStackTrace();
                    throw e;
          } finally {
               try {
                    if (cnx != null && !cnx.isClosed())
                         cnx.close();
               } catch (Exception e1) {
                    e1.printStackTrace();
     private UserTransaction getUserTransaction() throws NamingException {
          return (UserTransaction) new InitialContext().lookup("java:comp/UserTransaction");
     private Connection getConnection(String jndiName) throws NamingException, SQLException {
          DataSource ds = (DataSource) new InitialContext().lookup(jndiName);
          return ds.getConnection();
%>
<%
launchTest();
     SQL CODE :
     CREATE TABLE TEST_TABLE (     TEST_COL NUMBER(2));
     insert into test_table values(1);
     insert into test_table values(2);
     insert into test_table values(3);
     insert into test_table values(4);
     insert into test_table values(5);
     insert into test_table values(6);
     insert into test_table values(7);
     insert into test_table values(8);
     insert into test_table values(9);
     insert into test_table values(10);
     insert into test_table values(11);
%>
</pre>

Nicolas Mervaillie wrote:
Hi,
I have a problem accessing multiple XA datasources :
- launch an sql request on datasource 1
- rs.next() on the resultset
- use the data to launch another sql request on datasource 2
After 10 iterations, the next() method throws an SQLException (ORA-01002: fetch out of sequence).
After further investigation, I noticed that :
- the problem doesn't occur if the same datasource is used for the 2 requests
- if I set the fetch size to 15 for the first request, the exception is thrown after 15 iterations (but I can't use this as a workaround because the number is potentially above the million).
Anyone experiencing the same problem ?Hi. This is a known weakness of Oracle. If the XA transactional context of a connection
changes during a result set processing, even if it is switched back correctly, the result
set is aborted. If the oracle driver has fetched 10 rows, that's all you get. The next
call to next() that needs real DBMS communication will fail.
By spec, an XA connection should be able to be swapped out and enlisted in different
transactions at a per-call granularity, so our pools allow that by default. Try setting
the KeepXAConnTillTxCOmplete setting in your pool.
When a connection is put back in the pool, if it is an XA connection we will suspend
any user tx, and test it with our own test tx, then re-enlist and re-start the user
tx in flight. This may be the context switch that kills the oracle result set.
Joe
>
Thanks in advance
Nicolas
Here's my configuration :
- Weblogic 8.1 SP4
- JDK sun 1.4.2_04 (we have the same problem with various JDKs including JRockit)
- Oracle 10g
- Oracle thin driver xa 10.1.0.2.0
A JSP I use to reproduce the problem :
<pre><%@ page language="java" contentType="text/html; charset=ISO-8859-1" pageEncoding="ISO-8859-1"%>
<%@ page import="java.util.ArrayList,
                    java.sql.Connection,
                    java.sql.PreparedStatement,
                    java.sql.ResultSet,
                    java.sql.SQLException,
                    javax.naming.InitialContext,
                    javax.naming.NamingException,
                    javax.sql.DataSource,
                    javax.transaction.UserTransaction"%>
<%!
     public void launchTest() throws Exception {
          Connection cnx =null ;
          try {
               getUserTransaction().begin();
               // SQL access #1
               cnx = getConnection("jdbc/xaDatasource1");
               PreparedStatement stmt = cnx.prepareStatement("SELECT test_col from test_table");
               ResultSet rs = stmt.executeQuery();
               // iterate on resulset from SQL 1
               int i=1;
               while (rs.next()){ // SQL exception after 10 iterations
                    System.out.println(i + "");
                    i++;
                    // SQL access #2
                    Connection cnx2 = getConnection("jdbc/xaDatasource2");
                    // problem occurs even if we don't request
                    cnx2.close();
                    // end SQL access #2
               getUserTransaction().commit();
          } catch (Exception e) {
               e.printStackTrace();
               try {
                    getUserTransaction().rollback();
               } catch (Exception e1) {
                    e1.printStackTrace();
                    throw e;
          } finally {
               try {
                    if (cnx != null && !cnx.isClosed())
                         cnx.close();
               } catch (Exception e1) {
                    e1.printStackTrace();
     private UserTransaction getUserTransaction() throws NamingException {
          return (UserTransaction) new InitialContext().lookup("java:comp/UserTransaction");
     private Connection getConnection(String jndiName) throws NamingException, SQLException {
          DataSource ds = (DataSource) new InitialContext().lookup(jndiName);
          return ds.getConnection();
%>
<%
launchTest();
     SQL CODE :
     CREATE TABLE TEST_TABLE (     TEST_COL NUMBER(2));
     insert into test_table values(1);
     insert into test_table values(2);
     insert into test_table values(3);
     insert into test_table values(4);
     insert into test_table values(5);
     insert into test_table values(6);
     insert into test_table values(7);
     insert into test_table values(8);
     insert into test_table values(9);
     insert into test_table values(10);
     insert into test_table values(11);
%>
</pre>

Configuring Multiple LDAP Datasources in VDS

Hi,
I'm trying to configure multiple LDAP Datasources using VDS, one talking to AD and other to Novell eDir from VDS, my LDAP connection strings works well but when I start the service in VDS the service will never startup all I see is Exception null, it does not throw any exception at the same time it doesn't start up the service. I've tried configuring with signle Datasource which works fine. This is failing when I combine those two datasources into one configuration. Have any configured multiple datasources with in VDS. Not sure if you have encountered any problems.
Thanks,
Joe.P

Are you just trying to bring in two LDAP data sources or do a join between them?
Actually both I believe are considered types of joins.
You cannot just define two datasources and expect them to show up.

Packaged jdbc modules and multi datasources

Is it possible to configure a multi-datasource within a packaged jdbc module within an application ear? It appears to me that multi-datasources are only supported as system datasources.

You can configure Multi-datasources within a packaged jdbc module:-
example :-
<jdbc-data-source xmlns="http://www.bea.com/ns/weblogic/91">
<name>examples-demoXA-multi-data-source</name>
<jdbc-data-source-params>
<jndi-name>examples-demoXA -multi-data-source</jndi-name>
<algorithm-type>Load-Balancing</algorithm-type>
<data-source-list>examples-demoXA,examples-demoXA-2</data-source-list>
</jdbc-data-source-params>
</jdbc-data-source>
All data sources in a multi data source for an application-scoped JDBC module must also be application scoped.
Link : http://edocs/wls/docs92/jdbc_admin/packagedjdbc.html

Multiple multi state objects

Hello all,
I want to export the multiple Multi state objects from Indesign.
var mso_total_count = MSO.states.length - app.documents[0].multiStateObjects.length;
     for (var i = 1; i <= mso_total_count; i++)
            var imageName = imgName_mso+".png";
            imageSavePath = File( "~/Desktop/omnion_mw/"+nav_Project_Name+"/images/"+ imageName);
            MSO.states[i].pageItems[0].exportFile(ExportFormat.PNG_FORMAT, imageSavePath);
            var add_event = MSO.states[i].pageItems[0].addEventListener;
By using this i am getting images of only one multi state object. If i have multiple multi state objects are there like Muti-state 1, Multi-stste 2 etc it doesnt work.
Can anyone help me on this.

@piggychops – I'm not exactly sure, but by that code you only cycle through the states of the MSOs.
You never change its visible appearance on the page, because the active state of the MSO is always the same state.
So you'll end up with always the same contents that is exported to PNG (do you get that result?).
To get all states active, set the states in the loop (one after another) to the active one and then export to PNG.
MSO.states[i].active = true;
But before, store the current active state, so that you can return to the original appearance…
Something slightly different, but the code may help you (all comments are in German, but that should be no problem):
Multistateobject exportFile Problem
http://www.hilfdirselbst.ch/gforum/gforum.cgi?post=498745#498745
http://www.hilfdirselbst.ch/gforum/gforum.cgi?post=498942#498942
Uwe

Multiple abap datasource?

hello,
is it possible to have multiple abap datasource for authentificating users in the portal?
if so, how to do that?
best regards

Hi,
I would say that this is not possible. You can of course use a CUA in order to "span" multiple ABAP systems and access the CUA, but I don't think that you can modify the dataSourceConfiguration to access multiple ABAP systems (compared with the way you can do it with LDAP Directories)
Regards,
Holger.

WLST - Get List of JDBC Multi DataSource Only..

Hi All, I am trying to get a list of Multi DataSources using wlst. The list should only have multi data source not the regular ones. I am able to get the all but not the multi data sources, is there an attribute which can uniquely identify the Multi data source. Please advise.

This is a bit hacky but seems to work:
from com.bea.wli.sb.management.configuration import ALSBConfigurationMBean
from com.bea.wli.config import Ref
alsbCore = findService(ALSBConfigurationMBean.NAME, ALSBConfigurationMBean.TYPE)
refs = alsbCore.getRefs(Ref.DOMAIN)
it = refs.iterator()
while it.hasNext():
    r = it.next()
    if r.getTypeId() == Ref.PROJECT_REF:
        print(r.getProjectName())Edited by: paddys on Mar 5, 2012 4:48 PM
Code formatting

Cannot add multiple members of a failover cluster to a DFSR replication group

Server 2012 RTM. I have two physical servers, in two separate data centers 35 miles apart, with a GbE link over metro fibre between them. Both have a large (10TB+) local RAID storage arrays, but given the physical separation there is no physical shared storage.
The hosts need to be in a Windows failover cluster (WSFC), so that I can run high-availability VMs and SQL Availability Groups across these two hosts for HA and DR. VM and SQL app data storage is using a SOFS (scale out file server) network share on separate
servers.
I need to be able to use DFSR to replicate multi-TB user data file folders between the two local storage arrays on these two hosts for HA and DR. But when I try to add the second server to a DFSR replication group, I get the error:
The specified member is part of a failover cluster that is already a member of the replication group. You cannot add multiple members for the same cluster to a replication group.
I'm not clear why this has to be a restriction. I need to be able to replicate files somehow for HA & DR of the 10TB+ of file storage. I can't use a clustered file server for file storage, as I don't have any shared storage on these two servers. Likewise
I can't run a HA single DFSR target for the same reason (no shared storage) - and in any case, this doesn't solve the problem of replicating files between the two hosts for HA & DR. DFSR is the solution for replicating files storage across servers with
non-shared storage.
Why would there be a restriction against using DFSR between multiple hosts in a cluster, so long as you are not trying to replicate folders in a shared storage target accessible to both hosts (which would obviously be a problem)? So long as you are not replicating
folders in c:\ClusterStorage, there should be no conflict.
Is there a workaround or alternative solution?

Yes, I read that series. But it doesn't address the issue. The article is about making a DFSR target highly available. That won't help me here.
I need to be able to use DFSR to replicate files between two different servers, with those servers being in a WSFC for the purpose of providing other clustered services (Hyper-V, SQL availability groups, etc.). DFSR should not interfere with this, but it
is being blocked between nodes in the same WSFC for a reason that is not clear to me.
This is a valid use case and I can't see an alternative solution in the case where you only have two physical servers. Windows needs to be able to provide HA, DR, and replication of everything - VMs, SQL, and file folders. But it seems that this artificial
barrier is causing us to need to choose either clustered services or DFSR between nodes. But I can't see any rationale to block DFSR between cluster nodes - especially those without shared storage.
Perhaps this blanket block should be changed to a more selective block at the DFSR folder level, not the node level.

Support for Multiple Multi-Touch Screens

hello,
I have written some code that implements the multi-touch and it has been working fine with just one multi-touch screen. However, someone using my program with 2 multi-touch screens hooked up is getting no input from the second device. I did not
see anything documentation referring to more than one multi-touch device, although the existence of hSource in the TOUCHINPUT structure would imply multiple devices are supported. After searching the forums, I found this post
http://social.msdn.microsoft.com/Forums/en-US/tabletandtouch/thread/9c11871d-945c-4f7b-bdaa-a1b1452959b8/
in which someone stated that 2 multi-touch monitors should work fine, but not with simultaneous input. Since the post is almost 2 years old, I am wondering if anything has changed, and whether this might be the problem my user is having. Or is
there some way to tell if it is a driver issue? I am not using the gesture library, just retrieving data with GetTouchInputInfo. I'll appreciate if anybody can point me to documentation or mention their experiences with multiple devices.
thank you.

thank you for your quick reply. I have gotten a few more details and the problem is not the scenario you suspected. They are using 1 window spanning 2 multi-touch monitors. The problem does seem to be related to simultaneous input. If
they have a finger down on the first monitor, my program does not receive any events from the second monitor. If they are not pressing down on the first monitor, the input from the second monitor comes through.
Does Windows 7 support simultaneous input from 2 or more multi-touch monitors? Is there something I have to do so that my program will work correctly? Or are there driver specific issues, and if so, where can I get help for that?
thank you.

Multiple Data Source (for FAILOVER) in Easy Connect Naming method

Hi All,
I was wondering if there is a way to specify multiple addresses as part of data source property in the connection string when trying to connect to Oracle with Easy Connect Naming method? this is basically for failover scenario.
I know its straight forward, if i use Connect Descriptor. But not sure how to do it with EZConnect
For ex. this is what I want to achieve with EZConnect
sales.us.example.com=
(DESCRIPTION=
(ADDRESS_LIST=
(LOAD_BALANCE=off)
(FAILOVER=ON)
*(ADDRESS=(PROTOCOL=tcp)(HOST=sales1-server)(PORT=1521))*
*(ADDRESS=(PROTOCOL=tcp)(HOST=sales2-server)(PORT=1521)))*
(CONNECT_DATA=(SERVICE_NAME=sales.us.example.com)))
I appreciate your suggestions...
Thanks

You might get better answers in one of the DB forums from a generic fail-over perspective as EZconnect isn't specific to ODP.NET.
I don't know if there's a way to do it or not with EZConnect version of the connect string, but you can certainly provide the fully qualified connect string without using tnsnames.ora..
constring = "data source=(DESCRIPTION=(ADDRESS_LIST=(LOAD_BALANCE=off)(FAILOVER=ON)(ADDRESS=(PROTOCOL=tcp)(HOST=sales1-server)(PORT=1521))(ADDRESS=(PROTOCOL=tcp)(HOST=sales2-server)(PORT=1521)))(CONNECT_DATA=(SERVICE_NAME=sales.us.example.com))); user id=scott;password=tiger"
Hope it helps,
Gres

Using Windows Search for multiple, multi-page PDFs

I am scanning documents into a Windows folder (documents have been OCR'ed). An employee verifies that each document has a specific number that OCR picked up, if OCR did not, the user will use the 'Add or edit text box' under tools in Adobe Acrobat Professional XI to add the correct number. Multiple scans are handled this way and merged together to create documents that may be up to 2300 pages long (so far). Within that same Windows folder, there are other files with similarly handled information and numerous pages.
My problem is, if I am looking for a specific number that is within one of the PDF files within the Windows folder - I can find it if I open the file that contains the number and do a search. However, even though we are using a Windows index to be able to search from the Windows Search, the Windows Search will not always be able to find one of the indexed numbers. Does anyone know if there is a maximum size or number of pages a PDF can be for the Windows Search to be able to be used? Perhaps this is a question I should ask of Windows. I am not sure if I am doing something wrong with the PDFs or if the Windex index isn't working. Thoughts?

Something that may help -
With a multi-page PDF open in Acrobat Pro; select Document > Extract Pages
In the Extract Pages dialog you can select "Extract Pages As Separate Files".
This will provide you with a PDF file for each page of the source PDF file.
Additional options are also possible.
In a thread at AUC, Thom Parker identifies these.
http://www.acrobatusers.com/forums/aucbb/viewtopic.php?id=4703
Be well...

Multiple multi-datasource to failover

Similar Messages

Maybe you are looking for