Load balancing is not happening on RAC (oracle 11g Rel 2) on Linux 5.1 - ORDS, SODA & JSON in the Database

Load balancing is not happening on Oracle 11g RAC. We have defined the SCAN listener and confured the TNS accordingly. Still we get the load maximum on node2 and less load on Node1 or vice versa. Please suggest me what to do? 

This is not the right forum to ask this question.


Oracle Clusterware V/S Sun Cluster 3.x For Two node Oracle RAC Deployments

Hello Experts
I am trying to find out some info Like Pros/cons & Advantages and disadvantages of Deploying Two Node Oracle RAC using oracle's Clusterware & the same info for deployment of Oracle RAC using Sun Cluster 3.x .........appretiate your views on the same ........is it better to use oracle clusterware instead of Sun Cluster 3.x ......since the cluster products from oracle might be well integrated with Oracle's DataBase...& apart from that may be better you have single vendor supporting end to end rac deployment ???? any views .........pls share
Thanks in Advance
first I would like to point out, that it is not an either/or question. You cannot run Oracle 10g RAC with Oracle's clusterware. That was still possible with 9i, but not with 10g. So the question is: What are the benefits or disadvantages of running CRS/Clusterware with Sun Cluster 3.x. I can only see benefits in doing so.
- You combine the best of both worlds, a clustering solution that handles the DB components, that comes with the DB, and has been implemented by the DB vendor - combined with a clustering infrastructure that is tightly integrated into the OS, that has been implemented by the OS vendor
RAC will always run on an OS supported by another vendor; so you'll never have this end to end RAC experience.
- keep in mind that clusterware is a portable clustering solution; therefor it cannot be integrated well enough into all of the OSes it is designed to run on; that's what comes with Sun Cluster:
- failure fencing on the disks
- out of the box HA interconnects
- better and faster node failure detection
- single device namespace
- robust heartbeat
- etc.
There is a whitepaper coming soon on this topic. So stay tuned
I just wanted to try and clarify what Hartmut said. First, when I talk about Oracle Clusterware and RAC, it implies Oracle 10g and not 9i because Clusterware only became available with 10g.
Next, you can run Oracle RAC on Solaris/SPARC or Solaris/x86/x64 without Sun Cluster, i.e. just using Clusterware. However, I would recommend using Sun Cluster for all the reasons that Hartmut mentions. The whitepaper that I have authored, and to which Hartmut and others contributed, discusses these benefits in detail. All being well, the paper will be available soon.
Any update on when the paper will be available?
Thanks in advance 
Sorry, found it on the cluster page not the blueprint page.
Thanks again 
the paper is available at:
See the top entry in the spotlight section.

Upgrade Oracle 8.1.6 to 9i on Sun Cluster 2.2

Hi All,
I'm going to upgrade Oracle 8.1.6 running a sun cluster server with HA 2.2 to Oracle 9i. There are more than one instance running on the cluster and we have to upgrade all with minimum downtime. My questions are:
-Can Oracle 9i running on HA 2.2 ?The OS is Sun 2.7
-Can Oracle 8i and 9i instances running on the same cluster nodes ? I ask this question because if we can not upgrade all databases then we have to sun some on 9i and some on 8i until we complete the upgrade later
Thanks in advance,
the answers are yes and yes
there are a few caviats - make sure your SUNWscor package is patched to the correct level to ensure you can start things up correctly (sqlplus versus svrmgrl) - you can tell this by looking at /etc/opt/SUNWscor/haoracle_support and see if there's a 9.x.* entry in there.
I'm not sure if you're going to need one or two listeners (and I thought there were rules about that - you may need to start either one listener for each - using different ports - or have the later version listen for both instances) but I'd imagine that one of the more DB friendly people who watch this area will gladly chime in and clarify or correct my statements.
Thanks much for your reply
Do I need to rerun the scistall utility to add an entry for 9i or can do it manually ?
We are considering to upgrade HA2.2 to HA3.0 or 3.1. I intend to upgrade HA fisrt then Oracle but also think that upgrade Oracle first also an option. What is the best way to do please.
Hi Bao,
You should upgrade to Cluster 3.1 (10/04) first, then Oracle 9i. The reason is that Cluster 3.1 (10/04) provides better support for Oracle 9i. As the post before mentioned, Oracle 9i uses sqlplus to start the database instead of svrmgrl. The Oracle HA Agent on Cluster 3.1 already is Oracle 9i aware and already uses sqlplus to start the database. In Cluster 3.0 and 2.2, the Oracle HA agent uses svrmgrl to start the database. You would have to patch the cluster software or else starting Oracle9i with svrmgrl will fail because svrmgrl does not exist in Oracle 9i.
You can also support Oracle 8i and 9i on Cluster 3.0 and 3.1. You just install Oracle with the installer and modified the appropriate configuration files so that the cluster monitiors your Oracle 8i and 9i instances. You only need on listener if you want to use the standard port 1521. You would use Oracle 9i listener to also listen for Oracle 8i. However, you might want to use 2 listeners depending upon your situation.

Upgrade via failover

I have a 750Gb db that I need to upgrade to 12c.   I was wondering if I could create a standby database on a server running 12c and then switchover to the 12c database.   I've heard opinions saying yes and no so thought I'd pose the question to a larger audience.    I've looked at https://docs.oracle.com/cd/B19306_01/server.102/b14239/create_ps.htm#i70517 and it doesn't really say anything about oracle versions that I saw. 

Has Anyone configured Oracle GoldenGate for HA w/Oracle Clusterware 11gR2 ?

I am looking for some perspective of real world cases with respects to configuring Oracle GoldenGate 11.2 for HA with Oracle Clusterware
I have seen essentially two methods for this configuration:
<OL><LI>Using the Bundled Agents (XAG) with Clusterware , specifically the agent for GoldenGate</LI>
<LI>Using crsctl to add the GoldenGate instance process(s) as standard application cluster resources , configured with application action script(s).</LI></OL>
I have reviewed both options covered in these documents and either seem like they'd work fine.
Option 1 covered in this doc which came out in Sept 2012:
Option 2 covered in this doc which was released in March 2013:
Its curious to see that Option 2, which does not use the agents is the more recent doc.
Also, as some background information, I have a 2 node Grid Infrastructure cluster on OEL 6.3. On this 2 node cluster, I am running an single instance Oracle database configured for HA as an application cluster resource; similar to the configuration defined in this doc:
My plan is to implement GoldenGate Integrated capture on the same cluster..either on the same node as the database, or on the 2nd node as a Downstream Integrated Capture instance.
I can see some benefits of both options, and was curious to know if anyone has this type of configuration running in an active dev/test/prod environment.
If so, was there a decision point that made one option better than the other ?
Edited by: userTK421 on May 17, 2013 1:51 PM

MYSQL cluster & DR

Hi all, I am asked to setup MYSQL DR, in which there is already cluster is implemented. Now, they asked me to setup DR. totally New into MYSQL. Is there anyway to go? please ask any info if you need. Mysql 5.1 is implemented on cluster. Regards.Oracle DBA 
MySQL Cluster 5.1 is very old and the latest is 7.4 (see https://www.mysql.com/products/cluster/)  However many people use cluster in different ways and there are other options besides the MySQL Cluster product. Depending on what you need, there are other options.  I suggest reading the MySQL scalability guide (MySQL :: MySQL HA/Scalability Guide :: 1 High Availability and Scalability) as a starter.
You can consider MySQL Fabric for DR which provides Automatic Fail over.  If you prefer to scale your database  in future you can add sharding on top of it . But you can start using fabric just for DR without sharding. Mysql cluster will be more expensive solution and its uses NDB Storage engine. MySQL :: MySQL Fabric - A Guide to Managing MySQL High Availability & Scaling Out
Hi AustinTheDba, Can you please elaborate your question. MySQL Cluster  itself is a DR Solution and what is implementing DR on top of it ?  Thanks,DataBoy
MySQL Cluster is designed to be a shared nothing architecture.  If a server goes away, MySQL Cluster knows how to recover.  It can also be set up to replicate between data centers.  Please refer to https://www.mysql.com/products/cluster/availability.html