Re: Question on IO consideration

Thank You All. I was trying to query a few databases. And the below query gives Zero for most. But I checked all of them having the cell server version is 19.2+. Few of the databases are in 11.2 though. We do have big partition tables with compressed historical partitions as 'query high'. So I wanted to understand, as we are talking mainly three separate things

1)Database in memory option which needs additional license and also it will need inmemory_size parameter to be set as nonzero (which we don't have set currently). And required a minimum 12.1+ Db version and is independent of the cell server version. And I think it will also need specific changes to the table/column definition to make it in memory. Correct me if I'm wrong.

2)Columnar flash cache store format, which I believe is default for all 12.1+ cell server versions irrespective of database version. And oracle will do it automatically based on OLTP vs warehousing type load. So why am I seeing the results out of the below query as zero in some of our databases? And also 'cell physical IO bytes saved by columnar cache' as zero. Is there any other parameter that needs to be tweaked to use this feature?

3)And _at_Ghassan , as you mentioned just now, is 'in memory flash cache' . Is it different from the above two and is dependent on specific DB and cell server version or parameter change?

Apology if I am mixing up multiple things and interpreting them wrong.

select
xmlcast(xmlquery('/cell_stats/stats/stats/stat[_at_name="columnar_cache_size"]' passing xmltype(STATISTICS_VALUE) returning content) as varchar2(200) ) "COLUMNAR CACHE SIZE"
from v$cell_state
where xmlexists('/cell_stats/stats/stats[_at_type="columnarcache"]' passing xmltype(STATISTICS_VALUE));

On Wed, Sep 22, 2021 at 1:22 PM Ghassan Salem <salem.ghassan_at_gmail.com> wrote:

> Mladen,
> The feature Tanel is mentioning is free, not to be confused with in-memory
> flashcache, which requires the inmemory option.
>
> Regards
>
> On Wed, Sep 22, 2021 at 1:59 AM Mladen Gogala <gogala.mladen_at_gmail.com>
> wrote:
>
>>
>> On 9/21/21 17:41, Tanel Poder wrote:
>> > Exadata storage cells (starting from cellsrv v12.1.2.1.0 / Jan 2015)
>> > can use fully columnar flash cache for greatly speeding up reads (the
>> > cache is fully columnar, not hybrid like the datafile storage).
>>
>>
>> And all they want in exchange for that capability is your first-born.
>> Bargain price, indeed. BTW, MariaDB can also do that, a bit cheaper.
>>
>> --
>> Mladen Gogala
>> Database Consultant
>> Tel: (347) 321-1217
>> https://dbwhisperer.wordpress.com
>>
>> --
>> http://www.freelists.org/webpage/oracle-l
>>
>>
>>

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