Tag Archives: 12c

Adaptive serial direct path read decision ignores object statistics since 12.1

Posted on by Posted in 12c, adaptive serial direct path reads, oracle 1,876 Page views 1 Comment

On versions 11.2.0.2 – 11.2.0.4 Oracle uses objects’ statistics to make direct path reads decision (of course, if “_direct_read_decision_statistics_driven” haven’t changed it to “false”), and we can force serial direct reads on statement level using sql profiles with hints INDEX_STATS/TABLES_STATS, but since at least 12.1.0.2 this decision ignores statistics.

Btw, thanks to Jure Bratina, we know now, that we need to repeat hint TABLE_STATS at least twice to make it work 🙂 And from the following test case we know that it takes parameters from second one 🙂

Compare trace files of the following test cases:

table_stats2.sql

[sourcecode language=”sql”]
drop table t1;

create table t1 as select * from dual;
—————————-
pro ######################################;
exec dbms_stats.gather_table_stats(”,’T1′);
exec dbms_stats.set_table_stats(user,’T1′,numblks => 33333333);

col value new_val oldval noprint;
select value from v$statname n, v$mystat s
where n.statistic#=s.statistic# and n.name =’physical reads direct’;

—————————-
alter session set tracefile_identifier=’table_stats2′;
–alter session set events ‘10053 trace name context forever, level 1’;
alter session set events ‘trace[nsmtio] disk highest’;
select/*+ table_stats(t1, scale, blocks=66666666 rows=2222222222)
table_stats(t1, scale, blocks=99999999 rows=4444444444)
*/
*
from t1;
select value-&oldval directreads, value from v$statname n, v$mystat s
where n.statistic#=s.statistic# and n.name =’physical reads direct’;

[/sourcecode]

[collapse]

You can see that our hint successfully changed number of blocks and forced direct path reads on 11.2.0.4:
Oracle 11.2.0.4:

[sourcecode language=”sql” highlight=”3″]
NSMTIO: qertbFetch:DirectRead:[OBJECT_SIZE>VLOT]
NSMTIO: Additional Info: VLOT=797445
Object# = 78376, Object_Size = 66666666 blocks
SqlId = 7naawntkc57yx, plan_hash_value = 3617692013, Partition# = 0
[/sourcecode]

[collapse]

But on 12.1.0.2 and 12.2.0.1 we can see 2 lines with “NSMTIO: kcbism” with the different types(2 and 3) and different number of blocks, and the direct path read decision was based on second one – from segment header:

Oracle 12.1.0.2:

[sourcecode language=”sql” highlight=”1,2″]
NSMTIO: kcbism: islarge 1 next 0 nblks 66666666 type 3, bpid 65535, kcbisdbfc 0 kcbnhl 4096 kcbstt 983 keep_nb 0 kcbnbh 59010 kcbnwp 1
NSMTIO: kcbism: islarge 0 next 0 nblks 4 type 2, bpid 3, kcbisdbfc 0 kcbnhl 4096 kcbstt 983 keep_nb 0 kcbnbh 59010 kcbnwp 1
NSMTIO: qertbFetch:NoDirectRead:[- STT < OBJECT_SIZE < MTT]:Obect’s size: 4 (blocks), Threshold: MTT(4917 blocks),
_object_statistics: enabled, Sage: enabled,
Direct Read for serial qry: enabled(::::::), Ascending SCN table scan: FALSE
flashback_table_scan: FALSE, Row Versions Query: FALSE
SqlId: 7naawntkc57yx, plan_hash_value: 3617692013, Object#: 302342, Parition#: 0 DW_scan: disabled
[/sourcecode]

[collapse]

Oracle 12.2.0.1:

[sourcecode language=”sql” highlight=”1,2″]
NSMTIO: kcbism: islarge 1 next 0 nblks 66666666 type 3, bpid 65535, kcbisdbfc 0 kcbnhl 4096 kcbstt 1214 keep_nb 0 kcbnbh 45026 kcbnwp 1
NSMTIO: kcbism: islarge 0 next 0 nblks 4 type 2, bpid 3, kcbisdbfc 0 kcbnhl 4096 kcbstt 1214 keep_nb 0 kcbnbh 45026 kcbnwp 1
NSMTIO: qertbFetch:NoDirectRead:[- STT < OBJECT_SIZE < MTT]:Obect’s size: 4 (blocks), Threshold: MTT(6072 blocks),
_object_statistics: enabled, Sage: enabled,
Direct Read for serial qry: enabled(:::::::), Ascending SCN table scan: FALSE
flashback_table_scan: FALSE, Row Versions Query: FALSE
SqlId: 7naawntkc57yx, plan_hash_value: 3617692013, Object#: 174411, Parition#: 0 DW_scan: disabled
[/sourcecode]

[collapse]

And similar example, but for IFFS(index fast full scan):
nb: I set the number of index blocks using dbms_stats to 33333000 and hinted the query with 77777700

index_stats2.sql

[sourcecode language=”sql”]
drop table t2 purge;
ALTER SESSION SET optimizer_dynamic_sampling = 0;
ALTER SESSION SET "_optimizer_use_feedback" = FALSE;
ALTER SESSION SET optimizer_adaptive_features = FALSE;
ALTER SESSION SET optimizer_adaptive_plans=FALSE;

create table t2(x) as select level from dual connect by level<=1000;
create index t2_ix on t2(1,x,rpad(x,100));

begin
dbms_stats.gather_table_stats(”,’T2′,cascade => true);
dbms_stats.set_table_stats(user,’T2′ ,numblks => 33333333);
dbms_stats.set_index_stats(user,’T2_IX’,numlblks => 33333000);
end;
/

col value new_val oldval noprint;
select value from v$statname n, v$mystat s
where n.statistic#=s.statistic# and n.name =’physical reads direct’;

alter session set tracefile_identifier=’index_stats2′;
alter session set events ‘trace[nsmtio] disk highest’;
select/*+ index_stats(t2, t2_ix, scale, blocks=7777700)
index_ffs(t2 t2_ix)
dynamic_sampling(0)
*/
count(*) cnt2
from t2;
select value-&oldval directreads, value from v$statname n, v$mystat s
where n.statistic#=s.statistic# and n.name =’physical reads direct’;

disc;
[/sourcecode]

[collapse]

You can see that on 11.2.0.4 oracle gets number of blocks from the hint (7777700)
Oracle 11.2.0.4 - index_stats:

[sourcecode language=”sql” highlight=”2″]
NSMTIO: qerixFetchFastFullScan:DirectRead[OBJECT_SIZE > VLOT]:
NSMTIO: AdditionalInfo: Object_size: 7777700 (blocks), vlot=797445
SqlId=by2zv0k566hj5, plan_hash_value=3419274230,Object#=78375
[/sourcecode]

[collapse]

From the first line we can see that kcbism takes the hinted number of blocks, but later kcbivlo rewrites it with the number from segment header:

Oracle 12.1.0.2 - index_stats:

[sourcecode language=”sql” highlight=”2″]
NSMTIO: kcbism: islarge 1 next 0 nblks 7777700 type 3, bpid 65535, kcbisdbfc 0 kcbnhl 4096 kcbstt 983 keep_nb 0 kcbnbh 59010 kcbnwp 1
NSMTIO: kcbivlo: nblks 22 vlot 500 pnb 49175 kcbisdbfc 0 is_large 0
NSMTIO: qerixFetchFastFullScan:[MTT < OBJECT_SIZE < VLOT]:NSMTIO: AdditionalInfo: Object_size: 22 (blocks), vlot=245875
SqlId=by2zv0k566hj5, plan_hash_value=3419274230,Object#=302347
[/sourcecode]

[collapse]

Oracle 12.2.0.1 - index_stats:

[sourcecode language=”sql” highlight=”2″]
NSMTIO: kcbism: islarge 1 next 0 nblks 7777700 type 3, bpid 65535, kcbisdbfc 0 kcbnhl 4096 kcbstt 1214 keep_nb 0 kcbnbh 45026 kcbnwp 1
NSMTIO: kcbivlo: nblks 22 vlot 500 pnb 60729 kcbisdbfc 0 is_large 0
NSMTIO: qerixFetchFastFullScan:[MTT < OBJECT_SIZE < VLOT]:NSMTIO: AdditionalInfo: Object_size: 22 (blocks), vlot=303645
SqlId=by2zv0k566hj5, plan_hash_value=3419274230,Object#=174409
[/sourcecode]

[collapse]

So we have 2 options to force direct reads:
1. to execute alter session set “_serial_direct_read”=’ALWAYS’;
2. or to force parallel plan to get parallel direct path reads (we can do it with even with dop=1)

How to speed up slow unicode migration of a table with xmltype columns

Posted on by Posted in oracle, parallel, query optimizing, troubleshooting 2,298 Page views Leave a comment

Recently I have had an issue with slow unicode migration of the database upgraded from 10g to 12.1.0.2. The main problem was a table with xmltype: we spent about 4 hours for this table(~17GB) during test migration, though all other tables (~190GB) migrated just for about 20 minutes.
We used DMU(Database Migration Assistant for Unicode), and the root cause of the problem was update statement generated by DMU for this table:

update  /*+ PARALLEL(A,16)*/ "RRR"."T_XMLDATA" A  set A."SYS_NC00011$" = SYS_OP_CSCONV(A."SYS_NC00011$", 'AL16UTF16')

“SYS_NC00011$” was internal hidden CLOB column used to store XMLTYPE. As you can see DMU added PARALLEL hint, but though oracle can use parallel dml for xmltype since 12.1.0.1, we can’t use it because of its’ limitations:

Changes in Oracle Database 12c Release 1 (12.1.0.1) for Oracle XML DB

Parallel DML Support for XMLType
Support for parallel DML has been improved for XMLType storage model binary XML using SecureFiles LOBs. The performance and scalability have been improved for both CREATE TABLE AS SELECT and INSERT AS SELECT.

Restrictions on Parallel DML

Parallel DML can be done on tables with LOB columns provided the table is partitioned. However, intra-partition parallelism is not supported.

For non-partitioned tables with LOB columns, parallel INSERT operations are supported provided that the LOB columns are declared as SecureFiles LOBs. Parallel UPDATE, DELETE, and MERGE operations on such tables are not supported.

Btw, Oracle didn’t support parallel dml for xmltype on previous releases:

No Parallel DML for XMLType – DML operations on XMLType data are always performed in serial. Parallel DML is not supported for XMLType. (Parallel query and DDL are supported for XMLType.)

So I had to use manual parallelization:
1. Monitor “Convert application tables” step through “View Table Conversion progress” and press “Stop” button during conversion of this table.
2. Create table with ROWIDs of this table and split them into 16 groups:

create table tmp_rids as 
select rowid rid, ntile(16)over(order by rowid) grp 
from t_xmldata;

3. Execute 

ALTER SYSTEM SET EVENTS '22838 TRACE NAME CONTEXT LEVEL 1,FOREVER';

to avoid “ORA-22839: Direct updates on SYS_NC columns are disallowed”
4. Start 16 sessions and each of them have to update own part:

update t_xmldata A 
set A."SYS_NC00011$" = SYS_OP_CSCONV(A."SYS_NC00011$", 'AL16UTF16') 
where rowid in (select rid from tmp_rids where grp=&grp);
commit;

5. Disable event 22838:

ALTER SYSTEM SET EVENTS '22838 TRACE NAME CONTEXT OFF';

6. Open “View Table Conversion progress” window, click on this table and change “Retry” to “Skip” option for the update step.

This simple method allowed to make unicode migration about 16 times faster.

12c: Little test of “TABLE ACCESS INMEMORY FULL” with count stopkey

Posted on by Posted in inmemory 1,982 Page views 5 Comments

The table has 9M rows:

SQL> with function f return int is
  2       begin
  3          for r in (select value from v$mystat natural join v$statname where name like 'IM scan rows') loop
  4             dbms_output.put_line(r.value);
  5             return r.value;
  6          end loop;
  7       end;
  8  select f() from t_inmemory where rownum<=1
  9  ;
 10  /

       F()
----------
         0

1 row selected.

SQL> /

       F()
----------
    491436

1 row selected.

SQL> /

       F()
----------
    982872

1 row selected.
DDL and Plan

[sourcecode language=”sql”]
create table t_inmemory inmemory
as
with gen as (select 0 id from dual connect by level<=3e3)
select 0 n from gen,gen;

SQL_ID cpgrrfv9h6m52, child number 0
————————————-
with function f return int is begin for r in (select value
from v$mystat natural join v$statname where name like ‘IM scan rows’)
loop dbms_output.put_line(r.value); return
r.value; end loop; end; select f() from t_inmemory where
rownum<=1

Plan hash value: 3697881339

———————————————————————————-
| Id | Operation | Name | Rows | Cost (%CPU)| Time |
———————————————————————————-
| 0 | SELECT STATEMENT | | | 3 (100)| |
|* 1 | COUNT STOPKEY | | | | |
| 2 | TABLE ACCESS INMEMORY FULL| T_INMEMORY | 1 | 3 (0)| 00:00:01 |
———————————————————————————-

Predicate Information (identified by operation id):
—————————————————

1 – filter(ROWNUM<=1)
[/sourcecode]

[collapse]

Oracle 12c: behavior tests of the Inline functions, “Identities” and “defaults”

Posted on by Posted in 12c, curious, undocumented 2,934 Page views 4 Comments

I have done several minitests:
1. SQL and PL/SQL engines: which functions will be executed if there are two functions with same name as in SQL, as in PL/SQL (like “USER”, LPAD/RPAD, etc..)
– PL/SQL.

PL/SQL

[sourcecode language=”sql”]
SQL> @trace_on
Enter value for trace_identifier: inline
Enter value for level: 12
Tracing was enabled:

TRACEFILE_NAME
—————————————————————————-
/u01/app/oracle/diag/rdbms/xtsql/xtsql/trace/xtsql_ora_21599_inline.trc

SQL> with
2 function inline_user return varchar2 is
3 begin
4 return user;
5 end;
6 select
7 inline_user
8 from dual
9 /

INLINE_USER
——————————
XTENDER

1 row selected.

SQL> @trace_off
— unlike SQL’s "USER", PL/SQL function SYS.STANDARD.USER recursively executes "select user from sys.dual":
SQL> !grep USER /u01/app/oracle/diag/rdbms/xtsql/xtsql/trace/xtsql_ora_21599_inline.trc
SELECT USER FROM SYS.DUAL

SQL>
[/sourcecode]

[collapse]

2. Will there be any context switches if we call the inline functions which contain another pl/sql functions/procedures?
– Yes

Test 1

[sourcecode language=”sql”]
SQL> sho parameter max_string

NAME TYPE VALUE
———————————— ———— ——————————
max_string_size string STANDARD

SQL> @trace_pl_on

Session altered.

SQL> with
2 function blabla(p_str varchar2) return varchar2 is
3 begin
4 return lpad(p_str, 5000, ‘*’);
5 end;
6 select
7 length(blabla(dummy)) lpad_plsql
8 from dual;
9 /
from dual
*
ERROR at line 8:
ORA-06502: PL/SQL: numeric or value error: character string buffer too small
ORA-06512: at line 5

SQL> @trace_pl_last.sql

RUNID EVENT_SEQ EVENT_COMMENT EVENT_UNIT_OWNER EVENT_UNIT
———- ———- ——————————– —————— ———–
1 1 PL/SQL Trace Tool started
1 2 Trace flags changed
1 3 PL/SQL Virtual Machine started <anonymous>
1 4 PL/SQL Internal Call <anonymous>
1 5 PL/SQL Virtual Machine stopped

[/sourcecode]

[collapse]

Test 2

[sourcecode language=”sql”]
SQL> @trace_pl_on

Session altered.

SQL> create or replace function f_standalone(p varchar2) return varchar2 is
2 begin
3 return lpad(‘x’,3)||p;
4 end;
5 /

Function created.

SQL> with
2 function blabla(p_str varchar2) return varchar2 is
3 s varchar2(32767);
4 begin
5 s:= lpad(p_str, 100, ‘1’);
6 s:= s||s;
7 s:= s||lpad(p_str, 100, ‘3’);
8 s:= s||s;
9 s:= s||(1+10);
10 s:= f_standalone(s);
11 s:= f_standalone(s);
12 s:= f_standalone(s);
13 return s;
14 end;
15 select
16 length(blabla(dummy)) lpad_plsql
17 from dual
18 /

LPAD_PLSQL
———-
611

SQL> @trace_pl_last.sql

RUNID EVENT_SEQ EVENT_COMMENT EVENT_UNIT_OWNER EVENT_UNIT
———- ———- ——————————– —————– ————
2 1 PL/SQL Trace Tool started
2 2 Trace flags changed
2 3 PL/SQL Virtual Machine started <anonymous>
2 4 PL/SQL Internal Call <anonymous>
2 5 PL/SQL Virtual Machine stopped
2 6 PL/SQL Virtual Machine started <anonymous>
2 7 PL/SQL Virtual Machine started <anonymous>
2 8 PL/SQL Internal Call <anonymous>
2 9 PL/SQL Virtual Machine stopped
2 10 PL/SQL Virtual Machine stopped
2 11 PL/SQL Virtual Machine started <anonymous>
2 12 PL/SQL Virtual Machine started <anonymous>
2 13 PL/SQL Internal Call <anonymous>
2 14 PL/SQL Virtual Machine stopped
2 15 PL/SQL Virtual Machine stopped
2 16 PL/SQL Virtual Machine started <anonymous>
2 17 PL/SQL Internal Call <anonymous>
2 18 PL/SQL Internal Call <anonymous>
2 19 Procedure Call <anonymous>
2 20 PL/SQL Internal Call XTENDER F_STANDALONE
2 21 Return from procedure call XTENDER F_STANDALONE
2 22 Procedure Call <anonymous>
2 23 PL/SQL Internal Call XTENDER F_STANDALONE
2 24 Return from procedure call XTENDER F_STANDALONE
2 25 Procedure Call <anonymous>
2 26 PL/SQL Internal Call XTENDER F_STANDALONE
2 27 Return from procedure call XTENDER F_STANDALONE
2 28 PL/SQL Virtual Machine stopped

28 rows selected.
[/sourcecode]

[collapse]

Test 3

[sourcecode language=”sql”]
SQL> @trace_pl_on

Session altered.

SQL> with
2 function blabla(p_str varchar2) return varchar2 is
3 s varchar2(32767);
4 begin
5 s:= lpad(p_str, 100, ‘1’);
6 s:= s||s;
7 s:= s||lpad(p_str, 100, ‘3’);
8 s:= s||s;
9 s:= s||(1+10);
10 return s;
11 end;
12 select
13 length(blabla(dummy)) lpad_plsql
14 from dual
15 /

LPAD_PLSQL
———-
602

1 row selected.

SQL> @trace_pl_last.sql

RUNID EVENT_SEQ EVENT_COMMENT EVENT_UNIT_OWNER EVENT_UNIT
———- ———- ——————————– —————— ————
3 1 PL/SQL Trace Tool started
3 2 Trace flags changed
3 3 PL/SQL Virtual Machine started <anonymous>
3 4 PL/SQL Internal Call <anonymous>
3 5 PL/SQL Internal Call <anonymous>
3 6 PL/SQL Virtual Machine stopped

6 rows selected.

[/sourcecode]

[collapse]

3. How IDENTITY works?
For all identity columns Oracle creates a sequence with name like “ISEQ$$_XXX”, where “XXX” is the object_id of the table. All identities we can get through DBA_TAB_IDENTITY_COLS.
All Identity sequences:

select i.*
      ,tab.owner       tab_owner
      ,tab.object_name tab_name
      ,sq.object_name  sequence_name
from sys.idnseq$ i
    ,dba_objects tab
    ,dba_objects sq
where tab.object_id=i.obj#
  and sq.object_id = i.seqobj#

And we can see usage of this sequence in plans:

[sourcecode language=”sql”]
SQL_ID fn5tjw6hu0dtn, child number 0
————————————-
insert into xt_identity (description) values(‘1’)

Plan hash value: 3838626111

————————————————————————————————–
| Id | Operation | Name | Starts | Cost | A-Rows | A-Time | Buffers |
————————————————————————————————–
| 0 | INSERT STATEMENT | | 1 | 1 | 0 |00:00:00.01 | 35 |
| 1 | LOAD TABLE CONVENTIONAL | | 1 | | 0 |00:00:00.01 | 35 |
| 2 | SEQUENCE | ISEQ$$_91720 | 1 | | 1 |00:00:00.01 | 4 |
————————————————————————————————–
[/sourcecode]

[collapse]

4. When executes “default seq.nextval”?
Test

[sourcecode language=”sql”]
SQL> create sequence xt_sq1;
SQL> create sequence xt_sq2;
SQL> create table xt_default(
2 id1 int default xt_sq1.nextval
3 , pad varchar2(30)
4 , id2 int default xt_sq2.nextval
5 );

Table created.

SQL> insert into xt_default(pad) values(‘1’);

1 row created.

SQL> select xt_sq1.currval, xt_sq2.currval from dual;

CURRVAL CURRVAL
———- ———-
1 1

SQL> insert into xt_default(pad) values(1/0);
insert into xt_default(pad) values(1/0)
*
ERROR at line 1:
ORA-01476: divisor is equal to zero

SQL> select xt_sq1.currval, xt_sq2.currval from dual;

CURRVAL CURRVAL
———- ———-
2 2
[/sourcecode]

[collapse]

Oracle 12c: Inconsistency of Inline “with” functions

Posted on by Posted in 12c, oracle, PL/SQL 2,648 Page views 2 Comments

I was hoping that if inline “with” functions are in the query, so their results will be consistent with it (as operators), but unfortunately such functions returns also inconsistent results as standalone pl/sql functions. 

SQL> create table t as select 1 a from dual;
Table created.

SQL> declare
  2    j binary_integer;
  3  begin
  4    dbms_job.submit( j
  5                    ,'begin
  6                        for i in 1..10 loop
  7                          dbms_lock.sleep(1);
  8                          update t set a=a+1;
  9                          commit;
 10                        end loop;
 11                      end;'
 12                   );
 13    commit;
 14  end;
 15  /

PL/SQL procedure successfully completed.

SQL> with
  2     function f return int is
  3       res int;
  4     begin
  5       dbms_lock.sleep(1);
  6       select a into res from t;
  7       return res;
  8     end;
  9  select
 10     f
 11  from dual
 12  connect by level<=10;
 13  /

         F
----------
         1
         1
         1
         2
         3
         4
         5
         6
         7
         8

10 rows selected.

Interesting: Jonathan Lewis wrote that inline “deterministic” functions doesn’t use caching mechanism as standalone deterministic functions.