DRM机制导致RAC实例重启

相关alert日志：
Wed Mar 20 22:43:14 EAT 2013
LMS2 (ospid: 4179) is not heartbeating for 216 seconds.
LMON detects unhealthy receivers.
Please check LMON and DIAG trace files for detail.
Wed Mar 20 22:43:19 EAT 2013
LMON (ospid: 4154) is terminating the instance.
LMON: terminating instance due to error 481
Wed Mar 20 22:43:19 EAT 2013
System state dump is made for local instance
System State dumped to trace file /oracle/app/admin/ecard/bdump/ecard2_diag_4150.trc
Wed Mar 20 22:43:21 EAT 2013
Shutting down instance (abort)
License high water mark = 780
Wed Mar 20 22:43:24 EAT 2013
Instance terminated by LMON, pid = 4154
Wed Mar 20 22:43:26 EAT 2013
Instance terminated by USER, pid = 21762
Wed Mar 20 22:43:48 EAT 2013
Starting ORACLE instance (normal)

DRM本身需要消耗资源，且存在大量bug，因此建议通过修改隐含参数屏蔽该特性：
_gc_affinity_time=0
_gc_undo_affinity=FALSE
并重启使其生效，对于无法将整个集群关闭重启的环境，下面的调高阈值的方式最佳接近方案：
SQL> ALTER system SET "_gc_affinity_limit" = 1000000 scope = BOTH;
System altered.
SQL> ALTER system SET "_gc_affinity_minimum"=10000000 scope = BOTH;
System altered.