0 Replies Latest reply: May 13, 2014 8:38 AM by Jimbo RSS

    A Dirty Dozen of Queries on Server Pools


      Hi, after reading around the accompanying Oracle documentation ( probably the source of my confusion ! ), I have a number of queries on Oracle’s new 11gR2 Policy Managed Resources / Database and Server Pools


      Q1. I presume a Server Pool is a cluster wide object rather than a local node object ?


      Q2. How are cluster nodes / servers assigned to a Server Pool in the first place ? Does the DBA manually assign servers to the pool or does the Clusterware do it and on what basis ?


      Q3. Can a cluster resource be assigned to only one Server Pool at a time ?


      Q4. Can a Server be assigned to more than 1 Server Pool at a time ?


      Q5. Under what circumstances does a server / node automatically move ( or get moved by Clusterware ) from one Pool to another ?


      Q6. How does a node get into the FREE Server Pool ?


      Q7. I understand the use of IMPORTANCE with a Server Pool in determining which Server Pool gets an available server but when do these servers become available and get assigned to the different Pools ? Is it on startup of individual cluster resources ? What if 2 competing Server Pools have the same Importance setting – what happens ?


      Q8. How does MIN_SIZE and MAX_SIZE of a Server Pool work ? Once the number of Servers assigned to a Pool reaches MIN_SIZE what would cause any more to be added ?


      Q9. Is the idea with a Server Pool that you simply state via the CARDINALITY setting of a cluster resource, how many nodes you wish the cluster resource to run on ?


      Q10. Does this then mean that the clusterware could choose any of the nodes in your cluster ( up to the CARDINALITY nos ) to run the cluster server on i.e. if I had an 8 node cluster and a cluster resource with CARDINALITY of 3 and assigned to a Server Pool, then any 3 nodes of the cluster, within that Server Pool could be used to run the cluster resource on ?


      Q11. Is it therefore possible that a Server Pool may have more nodes in it than the Cardinality of the cluster resource is defined with ? If so what happens ? Does the cluster service just start on the number of nodes specified by the cardinality ? What determines which nodes in the Server Pool would the cluster resource be started on and which would not ?


      Q12. I presume a cluster resource cannot be assigned to a Server Pool where the Cardinality of the resource is > MAX_SIZE for the Server Pool ?



      Well it is surely a confusing topic, so any help greatly appreciated

      thanks Jim