11g some parameter is DISK_REPAIR_TIME (default is 3.6h), represents more than a prescribed disk offline time to time, Oracle will automatically put the disk to drop out, then you need to manually add it to the disk group. 12c, has added a parameter FAILGROUP_REPAIR_TIME (default value 24h), also said: If a FAILGROUP OFFLINE exceeds the allotted time, it will be DROP off.
The introduction of this parameter is very good. Now one machine, the following storage nodes are basically a storage node as a FAILGROUP, so there are a storage node to a power failure a FAILGROUP OFFLINE is possible. 24h time is basically enough.
note:
l ALTER DISKGROUP OFFLINE DISK DROP AFTER 5m ignores the parameter value.
The read-only parameter 2 NORMAL HIGH redundant disk group and effective.
Modify the parameter syntax:
SQL> ALTER DISKGROUP data SET ATTRIBUTE 'disk_repair_time' = '4.5h';
SQL> ALTER DISKGROUP data SET ATTRIBUTE 'FAILGROUP_REPAIR_TIME' = '270m';