Question about i/o
we are on Oracle 10g on Red Hat Linux x86_64.. a 2-node RAC
I want to create a report which is something like what one gets from v$event_histogram for wait event db file sequential read but instead of the buckets that are preset I want the buckets to be aceh one ms. In other words I want to know hw many reads were < 1ms, how many were between 1ms and 2ms , etc.
I think this can also be achieved if there were a way to time each block read during a session (something like what you see in a 10046 level 8 trace file)