Forgot to mention that the problem disappears immediately if we cause the sql in question to hard parse - either flushing the shared pool or adding a hint > _____________________________________________ > > We have a 4-node Linux x86-64 cluster running 10.2.0.4 (CRS, ASM and > DB). For the last one week we are running into a strange issue. > One of SQLs which usually executes in under a sec suddenly starts > running into db sequential and read by other session waits. > We have outline hints to make sure the execution plan is not changing > and to eliminate sql related issues. > > The problem occurs on only one node and so far it has it every node > randomly. > > Checked for hot-block contention. Came up with only one with TCH of 50 > with more than thousand calls. > > The read by other session and db sequential read p1, p2 values change > every second to point to different datafiles and blocks from user's > default TS. > > Any help in identifying the issue or pointing in the right direction > will be appreciated very much. I am working with OSS as well on this. > > Thank you, > Younus > > >