You can set the archiving destination values as follows in the initialization parameter file for either policy-managed or administrator-managed databases.
Set the SID.LOG_ARCH_DEST parameter for each instance using the SID designator, as shown in the following example:
sid1.LOG_ARCHIVE_DEST_1="LOCATION=/arc_dest_1" sid2.LOG_ARCHIVE_DEST_1="LOCATION=/arc_dest_2" sid3.LOG_ARCHIVE_DEST_1="LOCATION=/arc_dest_3"
For policy-managed databases, manually create a node and instance binding to ensure that sid1 always runs on the same node, as follows:
$ srvctl modify database -d mydb -n node1 -i sid1 $ srvctl modify database -d mydb -n node2 -i sid2 $ srvctl modify database -d mydb -n node3 -i sid3
The following list shows the possible archived redo log entries in the database control file. Note that any node can read archived redo logs from any of the threads, which must happen in order for the database to recover after a failure.
/arc_dest_1/log_1_1000_23435343.arc /arc_dest_2/log_1_1001_23452345.arc <- thread 1 archived in node2 /arc_dest_2/log_3_1563_23452345.arc <- thread 3 archived in node2 /arc_dest_1/log_2_753_23452345.arc <- thread 2 archived in node1 /arc_dest_2/log_2_754_23452345.arc /arc_dest_3/log_3_1564_23452345.arc