Re: Why does an online redolog keeps active after checkpoint completion?

"quarkman" <quarkman_at_myrealbox.com> wrote in message news:oprta1dwt1zkogxn_at_haydn...
> On Sat, 2 Aug 2003 20:37:42 +0200, Anton Buijs
<remove_aammbuijs_at_xs4all.nl>
> wrote:
>
> > |
> >
> > You are right. I varied the value of log_checkpoint_timeout in my
> > playground
> > database and the status active becomes inactive after a period of time
> > that
> > is related to the timeout period. Lgwr is choosing priorities as you
> > described. I get the impression it also depends on the load on the db.
My
> > tests where the only active sessions. When I do some (update) work the
> > active status disappears also much quicker.
> > Thanks.
>
> It should be pointed out, however, that setting L_C_TIMEOUT is not a
> particularly good idea. In 8i, you have L_C_INTERVAL and
> FAST_START_IO_TARGET to work with to make instance recovery time extremely
> predictable. And in 9i, you have FAST_START_MTTR_TARGET to use, which
> internally gets converted into settings for _IO_TARGET and _INTERVAL -but
> not for _TIMEOUT. Which gives the game away: _TIMEOUT is not particularly
> useful in limiting instance recovery times in a predictable fashion.
>
> In the absence of _TIMEOUT, behaviour is as I described earlier: another
> log switch will cause the original log's status to be updated.
>
> ~QM

If L_C_INTERVAL is set to a value less than the redo log file size (by either directly setting the parameter, or indirectly setting it via setting FAST_START_MTTR_TARGET as you mentioned), you'll get another checkpoint before the next log switch, in which case the original log's status in the control file might get updated before the next log switch.

Cheers,
Dave Received on Mon Aug 04 2003 - 01:31:56 CDT