Re: optimizer_mismatch and hash_match_failed
- From: Johan Eriksson <valpis@xxxxxxxxx>
- To: Tanel Poder <tanel@xxxxxxxxxxxxxx>
- Date: Thu, 18 Apr 2013 19:05:47 +0200
No, cursor_charing are going to be changed, I just tries to have the db
alive until they have changed their code and tested it extensively :)
We have a lot of problems with similiar (too many child cursor until the DB
stopped working due to that, over 65K of childs for one cursor...)
On Thu, Apr 18, 2013 at 7:02 PM, Tanel Poder <tanel@xxxxxxxxxxxxxx> wrote:
> I would not use cursor_sharing = SIMILAR at all, it's deprecated and in my
> experience has caused more problems than it solved. If you definitely do
> need Oracle to replace literals with binds for you, then use FORCE. The
> adaptive bind peeking & cardinality feedback should deal with the "unsafe"
> bind conditions well enough by now.
>
> --
> *Tanel Poder*
> Enkitec (The Exadata Experts)
> Training <http://blog.tanelpoder.com/seminar/> |
> Troubleshooting<http://blog.tanelpoder.com/>
> | Exadata<http://www.amazon.com/Expert-Oracle-Exadata-Apress/dp/1430233923>
> | Voicee App <http://voic.ee/>
>
>
>
> On Thu, Apr 18, 2013 at 8:00 PM, Tanel Poder <tanel@xxxxxxxxxxxxxx> wrote:
>
>> It's the MEMORY_TARGET / Automatic Memory Management, which changes your
>> PGA_AGGREGATE_TARGET regularly, resulting in some _smm_* parameter changes,
>> which happen to be part of the optimizer environment, which cause new hard
>> parses when some existing cursors are used again.
>>
>> The HASH_MATCH_FAILED shows up when you use cursor_sharing_similar with
>> "unsafe" bind variables - like where clauses with bind variables on columns
>> with histograms or bind variables in various range, like, between or <, >
>> conditions.
>>
>> --
>> *Tanel Poder*
>> Enkitec (The Exadata Experts)
>> Training <http://blog.tanelpoder.com/seminar/> |
>> Troubleshooting<http://blog.tanelpoder.com/>
>> | Exadata<http://www.amazon.com/Expert-Oracle-Exadata-Apress/dp/1430233923>
>> | Voicee App <http://voic.ee/>
>>
>>
>>
--
//www.freelists.org/webpage/oracle-l
Other related posts:
