We have a consultant that is proposing this option for us. All I can find at present is Oracle's 2 page flier on the configuration.
I want to know if any one has this set up either in Development or production and what are the advantages/ disadvantages of this. I am not sure if I want to be the first install in the country for this.
Any white papers would be most beneficial.
I don't. But, the disadvantage I can say is when the node goes down, there is no failover to other instance. Also, if there is any scheduled OS maintenance and system reboot is required, there should a database outage.
Oracle RAC One Node, which is described in the White Paper that you linked, ranges between a "full" Oracle RAC and a Single Instance Database with respect to the availability it can provide. In simple terms, Oracle RAC One Node is a fully supported failover solution from Oracle based on the Oracle only stack (no other software required). However, you can easily upgrade an Oracle RAC One Node to full RAC (online) when required and you will receive additional features such as OMotion. Last but not least, Oracle RAC One Node offers a price advantage over the full Oracle RAC solution, since it has some limitations.
Oracle RAC One Node as of now is only available on Linux and will require an 11.2 stack. Other than that, please, rest assured that you will not be the first countrywide implementing this solution.
On http://otn.oracle.com/rac you will find two more white papers (in addition to the data sheet that you already have) that explains Oracle RAC One Node in more detail: " [Oracle Real Application Clusters One Node 11g Release 2 Technical Overview|http://www.oracle.com/technology/products/database/clustering/pdf/twp_rac1nodev1.4.pdf] " AND " [Oracle Real Application Clusters One Node 11g Release 2 Users Guide|http://www.oracle.com/technology/products/database/clustering/pdf/ug_raconenode-2009.pdf] "
Hope that helps. Thanks,
RAC One Node = active/passive failover
RAC = active/active (all nodes active concurrently).
I implemented Active/Passive using 10gR2 using clusterware failover scripts. Upon failure of the active node, the second node would automatically start the database and the VIP for the node would also fail to the secondary node and it the vip that would use to connect to the database. Oracle took this concept and productized it.
There was one other note/white paper in MOS from ~ 5-6 years ago detailing how to configure the scripts for a/p setup that I cannot seem to find now.
Edited by: onedbguru on Nov 23, 2012 5:08 AM