Parallel query slave sessions not distributed among RAC nodes
In my two-node (8 cores in each node) 11.1.0.7 RAC environment, a
network service was created with LOAD_BALANCE=YES. Load balance was
tested successfully by manually starting multiple sqlplus connections
(sqlplus username/password@network_service). The sqlplus sessions were
distributed evenly to both nodes.
There is a large partitioned table with a parallel degree of 8 set at the table level. When a query ran against this table, 8 parallel slave processes were spawned. However, these slave processes were consistently spawned on only one node, where the parallel execution coordinator was running. In other words, the parallel slave process sessions were not distributed between the two nodes. By the way, there was absolutely no other user activities in the RAC database thoughout the time.
There is a large partitioned table with a parallel degree of 8 set at the table level. When a query ran against this table, 8 parallel slave processes were spawned. However, these slave processes were consistently spawned on only one node, where the parallel execution coordinator was running. In other words, the parallel slave process sessions were not distributed between the two nodes. By the way, there was absolutely no other user activities in the RAC database thoughout the time.
0