RAC Multiple Instance vs Service vs Serverpool
From: Sanjay Mishra <smishra_97_at_yahoo.com>
Date: Sun, 27 Jul 2014 13:29:38 -0700
Message-ID: <1406492978.15686.YahooMailNeo_at_web140703.mail.bf1.yahoo.com>
Hi Looking to review the architecture of some of the exiting RAC deployment in the enterprise where we have 4-6 node multiple RAC deployment using 11g R2 on Linux or AIX environment. Most of the environment has seen consolidation of several single non-rac database to these RAC environment where everyone has multiple instance running. So any environment consolidated to RAC environment has got database instance created on all Nodes. While doing some review saw some queries on various application in multiple database were working goon in non-rac vs RAC environment and moreover there is no such requirement to have instance on all nodes except High availability So want to check the expert views on the following and please coorect me where I am wrong in my assumptions. We can have the following options 1. Database Instance on each node of RAC and so 5 node RAC-A has 5 instance for Database A and so are all multiple database on this RAC-A 2. Using Services - Need careful planning and can create multiple service and can make these service preffered on one or group and available on other 3. Using 11g R2 Serverpool Now ofcourse, I provide some information for issue in using Multple Database instance which are linked to Interconnect, Master Block, Block/Node Hop to get required block on node and so Service might also have the same above issue when different service based connection on different node are running query against same block even using different service on different node of the cluster. Moreover if nodes goes down, Service will relocate to other survicing one based on configuration but stays there even the original node is back Serverpool is still dark as not clear that if a service srvA is running on two node ServerPool as singleton service and if that node goes down, it will relocate to second node and raise question like * What will happen and how fast is this relocation done automatically by Oracle * If the original node is back, does this singleton service comes back to original node or will stay on second node like ole Service concept until manually relocated to original or required node. * If in 4 node RAC, I had two serverpool of two node each and if any serverpool both node crashed, how we can make sure that it can take one of the node from second available Serverpool and bring all Database/services earlier running on Serverpol A live on ServerPool 2 TIA Sanjay
Date: Sun, 27 Jul 2014 13:29:38 -0700
Message-ID: <1406492978.15686.YahooMailNeo_at_web140703.mail.bf1.yahoo.com>
Hi Looking to review the architecture of some of the exiting RAC deployment in the enterprise where we have 4-6 node multiple RAC deployment using 11g R2 on Linux or AIX environment. Most of the environment has seen consolidation of several single non-rac database to these RAC environment where everyone has multiple instance running. So any environment consolidated to RAC environment has got database instance created on all Nodes. While doing some review saw some queries on various application in multiple database were working goon in non-rac vs RAC environment and moreover there is no such requirement to have instance on all nodes except High availability So want to check the expert views on the following and please coorect me where I am wrong in my assumptions. We can have the following options 1. Database Instance on each node of RAC and so 5 node RAC-A has 5 instance for Database A and so are all multiple database on this RAC-A 2. Using Services - Need careful planning and can create multiple service and can make these service preffered on one or group and available on other 3. Using 11g R2 Serverpool Now ofcourse, I provide some information for issue in using Multple Database instance which are linked to Interconnect, Master Block, Block/Node Hop to get required block on node and so Service might also have the same above issue when different service based connection on different node are running query against same block even using different service on different node of the cluster. Moreover if nodes goes down, Service will relocate to other survicing one based on configuration but stays there even the original node is back Serverpool is still dark as not clear that if a service srvA is running on two node ServerPool as singleton service and if that node goes down, it will relocate to second node and raise question like * What will happen and how fast is this relocation done automatically by Oracle * If the original node is back, does this singleton service comes back to original node or will stay on second node like ole Service concept until manually relocated to original or required node. * If in 4 node RAC, I had two serverpool of two node each and if any serverpool both node crashed, how we can make sure that it can take one of the node from second available Serverpool and bring all Database/services earlier running on Serverpol A live on ServerPool 2 TIA Sanjay
-- http://www.freelists.org/webpage/oracle-lReceived on Sun Jul 27 2014 - 22:29:38 CEST