Re: REDO log shifts are blocking for DB updates !!!

Shane Murray wrote:
>
> At a log switch (possibly at every chkpt, I can't remember) the SGA is
> flushed which takes time and resources. In order to minimize the
<snip>

Actually, the dirty blocks in the SGA are flushed at each checkpoint, and a log switch forces a checkpoint. In this case they may want to do frequent checkpoints to reduce the amount of time each one takes. And I thought our users had strict response-time requirements!

-- 
---
Allen Kirby			AT&T ITS Production Services
akirby_at_att.com			Alpharetta, GA.