Hi all.

We have 2 node RAC machine on HP/UX Itanium (10.1.0.4)

Sometimes we see some queries suffer from latch: cache buffer chain event. problem is mostly accessing same block concurrently.


When I take 10046 trace for the queries I always see negative values for P1 value of the wait event.


WAIT #9: nam='latch: cache buffers chains' ela= 1 p1=-4611686000696528624 p2=116 p3=0

when I change this value to hexadecimal (both negative and positive) to use as an input to hot blocks script ( metalink (163424.1)), zero rows returns every time from the script.

what should I do to find the hot blocks ??? Am I hitting a bug ?



--
A. Coskan GUNDOGAR

Oracle DBA

http://coskan.wordpress.com