Had a result_cache problem today.
If you didn’t know it already, DBA_HIST_ACTIVE_SESS_HISTORY is a real boon for getting a very good idea of a) when the problem started and b) how wide the impact was.
300 session were backing up on a latch free wait. Turns out the latch was the “Result Cache: Latch”.
The blocking session was doing a select from a view that used some result_cache function lookups.
The blocking session having been killed and given a couple of minutes to clean up, the blockage cleared in a instant.
Still trying to get to the bottom of the result cache issue – is there a serious result_cache flaw or have I just implemented it poorly?
Related info:
- http://www.pythian.com/blogs/594/oracle-11g-query-result-cache-and-the-rc-enqueue
- http://www.pythian.com/blogs/598/oracle-11g-query-result-cache-rc-latches
- http://www.pythian.com/blogs/1004/oracle-11g-result-cache-in-the-real-world
- http://www.pythian.com/blogs/660/does-oracle-11gs-result-cache-scale-poorly
1 response so far ↓
Result_cache blocking II « OraStory // November 13, 2008 at 2:52 pm |
[...] Contact ← Result_cache blocking [...]