TEST CASE:
========
conn sys as sysdba
alter system set resource_limit=true;
CREATE PROFILE support_profile LIMIT SESSIONS_PER_USER 6;
alter user <USERNAME> profile support_profile;
conn <USERNAME>/<PASSWORD>
SQL> alter session set "_px_trace"="high","compilation","high","execution","messaging","time";
Session altered.
SQL> ALTER SESSION SET EVENTS '10053 trace name context forever, level 10';
Session altered.
SQL> select /*+ PARALLEL(8) monitor*/ count(*) from <TABLE NAME>;
COUNT(*)
918843
SQL> SET LONG 1000000
SQL> SET LONGCHUNKSIZE 1000000
SQL> SET LINESIZE 1000
SQL> SET PAGESIZE 0
SQL> SELECT DBMS_SQLTUNE.report_sql_monitor(sql_id => NULL, type => 'TEXT',report_level=>'ALL') AS report FROM dual;
SQL Monitoring Report
SQL Text
select /*+ PARALLEL(8) monitor*/ count(*) from sales
Parallel Execution Details (DOP=8 , Servers Allocated=8)
SET LONG 1000000
SET LONGCHUNKSIZE 1000000
SET LINESIZE 1000
SET PAGESIZE 0
SELECT DBMS_SQLTUNE.report_sql_monitor(sql_id => NULL, type => 'TEXT',report_level=>'ALL') AS report FROM dual;
Database profile is used to set SESSION_PER_USER to 6. As a result you cannot have more than 6 sessions. If you try to open 7 sessions you get and ORA-02391 error.
This also causes that an SQL may not get the requested number of parallel processes.
CHANGES
CAUSE
Expected behaviour since Bug 5910422 is fixed.
SOLUTION
This is the expected behaviour in 10.2.0.4 and above (where Bug 5910422 is fixed). The sessions_per_user limit is enforced when the parallel workers are starting in the session.
The following demonstrates this behaviour on one session. Obviously if more than one sessions are open and parallel statement is running, its DOP will be reduced accordingly.
The example demonstrates the following steps on a query that has parallel plan with 1 worker group having 1 worker set. The query is requested to run with dop=8 by statement level parallel hint.
- No profile limit set, and the query runs with DOP 8, having 8 parallel workers.
- Then profile LIMIT SESSIONS_PER_USER is set to 6. Oracle wants to create 8 parallel workers, before the 8 parallel workers would join the QC, it recognizes SESSIONS_PER_USER is set to 6, so it releases 3 workers, and end up with 5 workers + the QC.
This demonstration was done on single instance. It may apply to RAC but the meaning of SESSIONS_PER_USER profile limit in multi instance environment is out of the scope of this article.
STEPS
conn sh
select /*+ PARALLEL(8) monitor*/ count(*) from T1;
SET LONG 1000000
SET LONGCHUNKSIZE 1000000
SET LINESIZE 1000
SET PAGESIZE 0
SELECT DBMS_SQLTUNE.report_sql_monitor(sql_id => NULL, type => 'TEXT',report_level=>'ALL') AS report FROM dual;
-- enable resource plan and resource limit
conn sys as sysdba
alter system set resource_limit=true;
CREATE PROFILE support_profile LIMIT SESSIONS_PER_USER 6;
alter user sh profile support_profile;
conn sh
alter session set "_px_trace"="high","compilation","high","execution","messaging","time";
ALTER SESSION SET EVENTS '10053 trace name context forever, level 10';
select /*+ PARALLEL(8) monitor*/ count(*) from T1;
SET LONG 1000000
SET LONGCHUNKSIZE 1000000
SET LINESIZE 1000
SET PAGESIZE 0
SELECT DBMS_SQLTUNE.report_sql_monitor(sql_id => NULL, type => 'TEXT',report_level=>'ALL') AS report FROM dual;
-- clean up
conn sys as sysdba
drop PROFILE support_profile cascade;
alter system set resource_limit=false;
ACTUAL RESULT
SH@dw23> select /*+ PARALLEL(8) monitor*/ count(*) from T1;
COUNT(*)
783327
SH@dw23> SET LONG 1000000
SH@dw23> SET LONGCHUNKSIZE 1000000
SH@dw23> SET LINESIZE 1000
SH@dw23> SET PAGESIZE 0
SH@dw23> SELECT DBMS_SQLTUNE.report_sql_monitor(sql_id => NULL, type => 'TEXT',report_level=>'ALL') AS report FROM dual;
SQL Monitoring Report
SQL Text
select /*+ PARALLEL(8) monitor*/ count(*) from T1
Global Information
Status : DONE (ALL ROWS)
Instance ID : 1
Session : SH (11:3833)
SQL ID : 0md7m3v2sh21s
SQL Execution ID : 16777216
Execution Started : 02/20/2014 11:53:14
First Refresh Time : 02/20/2014 11:53:14
Last Refresh Time : 02/20/2014 11:53:15
Duration : 1s
Module/Action : SQL*Plus/-
Service : ABCD.xx.xyz.com
Program : sqlplus.exe
Fetch Calls : 1
Global Stats
=========================================================================================
| Elapsed | Cpu | IO | Concurrency | Other | Fetch | Buffer | Read | Read |
| Time(s) | Time(s) | Waits(s) | Waits(s) | Waits(s) | Calls | Gets | Reqs | Bytes |
=========================================================================================
| 5.40 | 0.16 | 3.73 | 0.21 | 1.30 | 1 | 2065 | 318 | 11MB |
=========================================================================================
Parallel Execution Details (DOP=8 , Servers Allocated=8 )
==================================================================================================================================================
| Name | Type | Server# | Elapsed | Cpu | IO | Concurrency | Other | Buffer | Read | Read | Wait Events |
| | | | Time(s) | Time(s) | Waits(s) | Waits(s) | Waits(s) | Gets | Reqs | Bytes | (sample #) |
==================================================================================================================================================
| PX Coordinator | QC | | 0.79 | 0.01 | 0.24 | 0.21 | 0.33 | 58 | 18 | 144KB | db file parallel read (1) |
| p000 | Set 1 | 1 | 0.59 | 0.02 | 0.46 | | 0.11 | 281 | 33 | 2MB | db file sequential read (1) |
| p001 | Set 1 | 2 | 0.59 | 0.02 | 0.42 | | 0.14 | 258 | 54 | 2MB | direct path read (1) |
| p002 | Set 1 | 3 | 0.58 | 0.02 | 0.41 | | 0.15 | 418 | 63 | 1MB | direct path read (1) |
| p003 | Set 1 | 4 | 0.56 | 0.01 | 0.42 | | 0.13 | 204 | 28 | 968KB | direct path read (1) |
| p004 | Set 1 | 5 | 0.56 | 0.04 | 0.39 | | 0.14 | 286 | 38 | 1MB | db file sequential read (1) |
| p005 | Set 1 | 6 | 0.58 | 0.02 | 0.46 | | 0.10 | 181 | 26 | 1MB | direct path read (1) |
| p006 | Set 1 | 7 | 0.56 | 0.02 | 0.48 | | 0.07 | 203 | 32 | 1MB | direct path read (1) |
| p007 | Set 1 | 8 | 0.58 | 0.01 | 0.45 | | 0.12 | 176 | 26 | 1MB | db file sequential read (1) |
==================================================================================================================================================
SQL Plan Monitoring Details (Plan Hash Value=1311612494)
===============================================================================================================================================================
| Id | Operation | Name | Rows | Cost | Time | Start | Execs | Rows | Read | Read | Activity | Activity Detail |
| | | | (Estim) | | Active(s) | Active | | (Actual) | Reqs | Bytes | (%) | (# samples) |
===============================================================================================================================================================
| 0 | SELECT STATEMENT | | | | 1 | +1 | 1 | 1 | | | | |
| 1 | SORT AGGREGATE | | 1 | | 1 | +1 | 1 | 1 | | | | |
| 2 | PX COORDINATOR | | | | 2 | +0 | 9 | 8 | 18 | 144KB | 11.11 | db file parallel read (1) |
| 3 | PX SEND QC (RANDOM) | :TQ10000 | 1 | | 1 | +1 | 8 | 8 | | | | |
| 4 | SORT AGGREGATE | | 1 | | 1 | +1 | 8 | 8 | | | | |
| 5 | VIEW | VW_TE_2 | 611K | 62 | 1 | +1 | 8 | 783K | | | | |
| 6 | UNION-ALL | | | | 1 | +1 | 8 | 783K | | | | |
| 7 | PX BLOCK ITERATOR | | 501K | 54 | 1 | +1 | 8 | 672K | | | | |
| 8 | TABLE ACCESS FULL | T1| 501K | 54 | 1 | +1 | 103 | 672K | 103 | 9MB | 55.56 | direct path read (5) |
| 9 | PX BLOCK ITERATOR | | 109K | 9 | 1 | +1 | 8 | 111K | | | | |
| 10 | TABLE ACCESS FULL | T1| 109K | 9 | 1 | +1 | 197 | 111K | 197 | 2MB | 33.33 | db file sequential read (3) |
===============================================================================================================================================================
SH@ABCD> conn sys/manager1@ABCD as sysdba
Connected.
SYS@ABCD> alter system set resource_limit=true;
System altered.
SYS@ABCD> CREATE PROFILE support_profile LIMIT SESSIONS_PER_USER 6;
Profile created.
SYS@ABCD> alter user sh profile support_profile;
User altered.
SYS@ABCD> conn sh/sh@ABCD
Connected.
SH@ABCD> alter session set "_px_trace"="high","compilation","high","execution","messaging","time";
Session altered.
SH@ABCD> ALTER SESSION SET EVENTS '10053 trace name context forever, level 10';
Session altered.
SH@dw23> select /*+ PARALLEL(8) monitor*/ count(*) from sales;
COUNT(*)
783327
SH@ABCD> SET LONG 1000000
SH@ABCD> SET LONGCHUNKSIZE 1000000
SH@ABCD> SET LINESIZE 1000
SH@ABCD> SET PAGESIZE 0
SH@ABCD> SELECT DBMS_SQLTUNE.report_sql_monitor(sql_id => NULL, type => 'TEXT',report_level=>'ALL') AS report FROM dual;
SQL Monitoring Report
SQL Text
select /*+ PARALLEL(8) monitor*/ count(*) from T1
Global Information
Status : DONE (ALL ROWS)
Instance ID : 1
Session : SH (11:3841)
SQL ID : 0md7m3v2sh21s
SQL Execution ID : 16777218
Execution Started : 02/20/2014 12:04:31
First Refresh Time : 02/20/2014 12:04:31
Last Refresh Time : 02/20/2014 12:04:32
Duration : 1s
Module/Action : SQL*Plus/-
Service : ABCD.xx.xyz.com
Program : sqlplus.exe
DOP Downgrade : 38%
Fetch Calls : 1
Global Stats
=========================================================================================
| Elapsed | Cpu | IO | Concurrency | Other | Fetch | Buffer | Read | Read |
| Time(s) | Time(s) | Waits(s) | Waits(s) | Waits(s) | Calls | Gets | Reqs | Bytes |
=========================================================================================
| 0.91 | 0.40 | 0.00 | 0.23 | 0.28 | 1 | 1618 | 62 | 9MB |
=========================================================================================
Parallel Execution Details (DOP=5 , Servers Requested=8 , Servers Allocated=5 )
==================================================================================================================================
| Name | Type | Server# | Elapsed | Cpu | IO | Concurrency | Other | Buffer | Read | Read | Wait Events |
| | | | Time(s) | Time(s) | Waits(s) | Waits(s) | Waits(s) | Gets | Reqs | Bytes | (sample #) |
==================================================================================================================================
| PX Coordinator | QC | | 0.35 | 0.06 | | 0.23 | 0.06 | 58 | | . | |
| p000 | Set 1 | 1 | 0.10 | 0.06 | 0.00 | | 0.04 | 395 | 16 | 2MB | |
| p001 | Set 1 | 2 | 0.10 | 0.10 | 0.00 | | 0.01 | 409 | 15 | 2MB | |
| p002 | Set 1 | 3 | 0.14 | 0.06 | 0.00 | | 0.07 | 281 | 10 | 2MB | |
| p003 | Set 1 | 4 | 0.12 | 0.09 | 0.00 | | 0.03 | 329 | 15 | 2MB | |
| p005 | Set 1 | 5 | 0.10 | 0.03 | 0.00 | | 0.07 | 146 | 6 | 872KB | |
==================================================================================================================================
SQL Plan Monitoring Details (Plan Hash Value=1311612494)
===================================================================================================================================================
| Id | Operation | Name | Rows | Cost | Time | Start | Execs | Rows | Read | Read | Activity | Activity Detail |
| | | | (Estim) | | Active(s) | Active | | (Actual) | Reqs | Bytes | (%) | (# samples) |
===================================================================================================================================================
| 0 | SELECT STATEMENT | | | | 1 | +1 | 1 | 1 | | | | |
| 1 | SORT AGGREGATE | | 1 | | 1 | +1 | 1 | 1 | | | | |
| 2 | PX COORDINATOR | | | | 1 | +1 | 6 | 5 | | | | |
| 3 | PX SEND QC (RANDOM) | :TQ10000 | 1 | | 1 | +1 | 5 | 5 | | | | |
| 4 | SORT AGGREGATE | | 1 | | 1 | +1 | 5 | 5 | | | | |
| 5 | VIEW | VW_TE_2 | 611K | 62 | 1 | +1 | 5 | 783K | | | | |
| 6 | UNION-ALL | | | | 1 | +1 | 5 | 783K | | | | |
| 7 | PX BLOCK ITERATOR | | 501K | 54 | 1 | +1 | 5 | 672K | | | | |
| 8 | TABLE ACCESS FULL | T1| 501K | 54 | 1 | +1 | 62 | 672K | 62 | 9MB | | |
| 9 | PX BLOCK ITERATOR | | 109K | 9 | 1 | +1 | 5 | 111K | | | | |
| 10 | TABLE ACCESS FULL | T1| 109K | 9 | 1 | +1 | 50 | 111K | | | | |
===================================================================================================================================================
PX_TRACE FILE SHOWS THE FOLLOWING
Oracle wants to create 8 parallel workers, before the 8 parallel workers would join the QC,
it recognizes SESSIONS_PER_USER 日志中没有显示 is set to 6, so it releases 3 workers, and end up with 5 workers + the QC.
kxfrAllocSlaves 70/ 0
DOP trace -- call kxfpgsg to get 8 workers
2014-02-20 12:04:32.414436 :PX_Messaging:kxfp.c@9813:kxfpgsg(begin):
reqthreads=8 height=0 lsize=0 alloc_flg=0x234
2014-02-20 12:04:32.414557 :PX_Messaging:kxfp.c@9886:kxfpgsg():
reqthreads=8 KXFPTHIS/KXFPLDBL/KXFPADPT/ load balancing=on adaptive=on
2014-02-20 12:04:32.414588 :PX_Messaging:kxfp.c@18418:kxfpiinfo():
inst cpus:mxslv:#pg
1 4 :270 :1
(default: 0) inst target is 270
number of active slaves on the instance: 0,
number of active slaves but available to use: 0
2014-02-20 12:04:32.415645 :PX_Messaging:kxfp.c@18742:kxfpclinfo():
inst(load :user:pct:fact:servtarget:queued:started:granted:active:active(free)): aff
1 (0.00:0 :100:400 :270 :0 :0 :0 :0 :0 )
defDOP=8, tpc=2
2014-02-20 12:04:32.415750 :PX_Messaging:kxfp.c@17831:kxfpLoadAdDOP():
granted dop: 8
load adapt num servers requested to = 8 (from kxfpAdaptDOP())
2014-02-20 12:04:32.415798 :PX_Messaging:kxfp.c@2149:kxfpqialo():
Allocate new q for dp=(nil)
new q=0x951b6f30 dp=(nil) pnum=65535 numa#=0 qser=0 done.
2014-02-20 12:04:32.415880 :PX_Messaging:kxfp.c@10255:kxfpgsg():
getting 1 sets of 8 threads, client parallel query execution flg=0x234
Height=8, Affinity List Size=0, inst_total=1, coord=1
Insts: 1
Threads: 8
getting 8 slaves (8 required) for q=0x951b6f30 qser=3073
...
2014-02-20 12:04:32.663551 :PX_Messaging:kxfp.c@3782:kxfpqsrls():
Release Slave q=0x951b6f30 qr=0x951b4f48 action=1 server=4 inst=1
...
2014-02-20 12:04:32.663853 :PX_Messaging:kxfp.c@3782:kxfpqsrls():
Release Slave q=0x951b6f30 qr=0x951b5758 action=1 server=6 inst=1
...
2014-02-20 12:04:32.664908 :PX_Messaging:kxfp.c@3782:kxfpqsrls():
Release Slave q=0x951b6f30 qr=0x951b4488 action=1 server=7 inst=1
...
2014-02-20 12:04:32.664939 :PX_Messaging:kxfp.c@11410:kxfpg1sg():
got 5 servers (sync), errors=0x0 returning
Acquired 5 slaves on 1 instances avg height=5 #set=1 qser=3073
P000 inst 1 spid 19450
P001 inst 1 spid 19452
P002 inst 1 spid 19454
P003 inst 1 spid 19456
P005 inst 1 spid 19458
2024-08-25 13:53:29.465 :PX_Messaging:kxfp.c@13629:kxfpg1sg():
got 4 servers (sync), errors=0x0 returning
2024-08-25 13:53:29.465 :PX_Messaging:kxfp.c@12144:kxfpgsg(): threads reduced from 8 to 4 due to not enough slaves
reason_code = 354, reason: DOP downgrade due to insufficient number of processes
Acquired 4 slaves on 1 instances avg height=4 #set=1 qser=6657
P006 set 1 inst 1 spid 22426
P001 set 1 inst 1 spid 6258
P004 set 1 inst 1 spid 22422
P003 set 1 inst 1 spid 6632
2024-08-25 13:53:29.465 :PX_Messaging:kxfp.c@12430:kxfpgsg():