Re: Sql net msg from client and fetch arrays

I've just changed it dynamically on a 19.3 instance, and now I'm getting 50 extra rows per second and with my startup parameters that's giving me about 9 minutes of v$active_session_history before data disappears off the "end" of the cyclic buffer.

Enterprise express does (at least in some places, I haven't checked everything) seem to understand how to report Average Active sessions; though the AWR report section on ASH manages to show all the top sessions and top events as things to do with Idle waits so it's not factoring out "real" active sessions.

Regards
Jonathan Lewis

On Mon, 25 Jan 2021 at 18:53, Tanel Poder <tanel_at_tanelpoder.com> wrote:

> Also, it would be interesting to see if the "Average Active Sessions"
> metric will just count all logged in sessions instead of active ones after
> changing this setting. And what 2nd order tasks, like monitoring alerts,
> maintenance, SQL tuning advisor, resource manager & etc decisions may go
> wrong as a result.
>
> --
> Tanel Poder
> https://tanelpoder.com/events
>
> On Mon, Jan 25, 2021 at 5:00 AM Jonathan Lewis <jlewisoracle_at_gmail.com>
> wrote:
>
>>
>> There is a hidden parameter "_ash_sample_all" with the obvious effect.
>> Defaults to FALSE, can be set to TRUE at the system level without
>> restarting the instance.
>> It's probably not a good idea for most systems though. I've seen so many
>> AWR reports showing connected sessions in the hundreds or thousands, with
>> active sessions running in the order of a few 10s.
>>
>>

--
http://www.freelists.org/webpage/oracle-l