Jonathan,
Would it make sense to also include die typecheck memory (TCHK) when measuring
the total shared memory consumed by an SQL?
For example, in Lothar’s case it was the TCHK allocation that failed.
In some cases, like “create table as select”, the TCHK shared memory can be
significantly larger than the SQLA part:
https://nenadnoveljic.com/blog/large-tchk-allocations-shared-pool/
Best regards,
Nenad
From: oracle-l-bounce@xxxxxxxxxxxxx <oracle-l-bounce@xxxxxxxxxxxxx> On Behalf
Of Jonathan Lewis
Sent: Dienstag, 18. Januar 2022 11:56
To: Lothar Flatz <l.flatz@xxxxxxxxxx>
Cc: Stefan Koehler <contact@xxxxxxxx>; oracle-l@xxxxxxxxxxxxx
Subject: Re: Long Parse Time for a big Statement
*** E-Mail from outside Vontobel: Do not click on links or open attachments
unless you know the content is safe. ***
I've got a note about the memory issue that suggests:
1) explain plan will use SGA memory while parsing information
2) executing the query to get the parsing done will use PGA
In my test case (using 19.11.0.0), executing the query demanded 313MB of PGA;
running explain plan for the query used only 47MB
At the same time the only significant changes in v$sgastat for explain vs.
execute where
shared pool free memory 470,539,784 Execute
shared pool free memory 217,137,088 Explain
shared pool SQLA 40,554,888 Execute
shared pool SQLA 295,652,112 Explain
In other words: explaining the plan demanded 250MB of memory from the shared
pool to use for SQLA
Regards
Jonathan Lewis
On Tue, 18 Jan 2022 at 09:11, Lothar Flatz
<l.flatz@xxxxxxxxxx<mailto:l.flatz@xxxxxxxxxx>> wrote:
Hi Stefan,
1.) No, I have not, I would have if I had access to the DB server. I did
construct a test case but right now I need to find a server with enough
memory to test it.
2.) Yep, there is both. Shared_pool seems to run out first.
Event waited on Times Max. Wait Total
Waited
---------------------------------------- Waited ----------
------------
PGA memory operation 147 0.00 0.00
latch: shared pool 9 0.00 0.00
SGA: allocation forcing component growth 27 2.51 2.51
..
ORA-04031: 80 Byte des Shared Memorys konnten nicht zugewiesen werden
("shared pool","explain plan for SELECT *
FR...","TCHK^4b493f9a","logdef: qcopCreateLog")
Thanks
Lothar
Am 18.01.2022 um 10:00 schrieb Stefan Koehler:
Hello Lothar,
sorry, if I missed some already explained details in the reply chain.
1) Have you already sampled the parsing (let's say with 20 CPU samples per
second for 2 or 3 minutes or so) and created a flame graph afterwards?
2) Are you sure that you are running out of memory in shared pool and not
also in PGA? Parsing allocates memory in both areas depending on what is
happening.
In the past I almost figured out every long parsing problem with CPU samples
and flame graphs (if PARSE timings were not available) - so please have a
look at the whole C-stack if you are just concerned about the parsing time
but if you are more concerned about the memory usage filter on kghalf, kghalo
and kghalp while sampling the C-stacks.
Best Regards
Stefan Koehler
Independent Oracle performance consultant and researcher
Website: http://www.soocs.de
Twitter: @OracleSK<
Lothar Flatz <l.flatz@xxxxxxxxxx<mailto:l.flatz@xxxxxxxxxx>> hat am--
17.01.2022 13:59 geschrieben:
Hi,
At one customer site we see generated statements, actually reports. The
parsetime for such a statement is over an hour, if it finishes at all.
It is possible we see "ORA-04031:" when we run out of memory in the
shared pool.
How big these statements are is hard to tell, since it depends on
formatting. With sql developer formatting i get in one typical example >
130000 lines.
The statements are constructed relatively simple.
It seems to be a kind of change report where columns from different
tables are retrieved.
At the beginning is a big case statement where a meaningful name is
generated for a value followed by this values. I counted 7400 case
entries as per statement in one case.
I addition we have a number of big inlists.
All this is running against a union view of 55 Tables.
In other words: If i want to stress the parser I would construct a
statement exactly like this.
However, one hour seems to be a unrealistically long parse time.
Even though that statement needs to be rewritten, but this will take time.
I want to know if there is any quick fix like increasing the shared pool
a lot. (Which I can't test unfortunately any time soon due lack of memory).
Any ideas how to speed up the parse time?
Database version is 19.7. Shared Pool size is 20GB
Thanks
Lothar
//www.freelists.org/webpage/oracle-l
