Oracle HANG by the MMAN spinning consuming a lot of CPU

某个库10.2.0.5的库MMAN进程占用100%的CPU导致数据库实例HANG。以下是metalink[ID 8211733.8]对此BUG 8211733的描述及解决方法。

Description

Shared pool latch contention high when shared pool is shrinking

Rediscovery Notes:

 Long shared pool latch wait times with call stack traces containing any of the following :

- kghquiesce_regular_extent

- kghquiesceable_extent_chkpin

- kgh_next_free

The latch may be held in kgh: quiesce ... by the MMAN spinning consuming a lot of CPU.

Workaround

 Disable shared pool durations.

 Set init.ora parameter _enable_shared_pool_durations=false

References

Bug:8211733 (This link will only work for PUBLISHED bugs)
Note:245840.1 Information on the sections in this article