G
George Potemkin
Guest
Richard, > For #4, a latch is required but I guess it could be a different latch as opposed to the USR latch. Spin latch instead of a queued latch? My question was not about optimization of the current statement cache mechanism. My point is that enabling type 2 statement caching is like keeping the light and air conditioning on all the time in a room that you are going to visit only once per week. It's just not efficient use of the resources (no matter expensive or not). "One Time" (type 3) statement caching would be better. Tom, > None the less -- a lot of times the "bad actor" is doing the same thing over and over and over. I guess it depends from application. I saw the examples when a few snapshots were needed to catch the "bad actor".
Continue reading...
Continue reading...