(HOWTO Oracle DB) - Performance problems due Un-Balanced CPU RAC nodes

Background:
Have you seen your production RAC database using lots of CPU in 1 node and NONE on other nodes?  Node1 seems to be particularly busy comparing to Node2, LAVG reaches 40+ in Node1, comparing to LAVG 2+ in Node2.


Question is Why?

This is because the database does not consider CPU utilization at the time of connecting a session. As result we have seen unbalanced CPU utilization across nodes in the cluster. In order to fix this problem, Oracle provides the ability to change the RAC load balancing algorithm. This is done by configuring the default service to use two load balancing features “Load Balancing Advisory” (LBA) and “Connection Load Balancing” (CLB). This change can be done on the fly with using the API “dbms_service.modify_service”.

According to the Oracle documentation DBMS_SERVICE uses the following parameters to modify load balancing mechanism.

DBMS_SERVICE.MODIFY_SERVICE(
   service_name        IN VARCHAR2,
   goal                IN NUMBER DEFAULT NULL,
   dtp                 IN BOOLEAN DEFAULT NULL,
   aq_ha_notifications IN BOOLEAN DEFAULT NULL,
   failover_method     IN VARCHAR2 DEFAULT NULL,
   failover_type       IN VARCHAR2 DEFAULT NULL,
   failover_retries    IN NUMBER DEFAULT NULL,
   failover_delay      IN NUMBER DEFAULT NULL,
   clb_goal            IN NUMBER DEFAULT NULL);

GOAL_NONE = 0 (Disables Load Balancing Advisory)

GOAL_SERVICE_TIME = 1 (Load Balancing Advisory is based on elapsed time for work done in the service plus available bandwidth to the service)
 
GOAL_THROUGHPUT = 2 (Load Balancing Advisory is based on the rate that work is completed in the service plus available bandwidth to the service)

CLB_GOAL_SHORT = 1 (Connection load balancing uses Load Balancing Advisory, when Load Balancing Advisory is enabled (either goal_service_time or goal_throughput). When GOAL=NONE (no load balancing advisory), connection load balancing uses an abridged advice based on CPU utilizatio)

CLB_GOAL_LONG = 2 (Balances the number of connections per instance using session count per service. This setting is recommended for applications with long connections such as forms. This setting can be used with Load Balancing Advisory when the connection pool is sized to accommodate gravitation within the pool itself (without adding or removing connections). The latter is the most efficient design).

Recommendation:
 In order to use CPU load balancing default service need to be changes to consider CPU connection load balancing on CLB parameter (set value to CLB_GOAL_SHORT or 1), and also set the load balancing goal to throughput (set value to GOAL_THROUGHPUT or 2).

Current setup is not using CLB

SQL> col name for a25
SQL> select inst_id,name,goal,clb_goal from gv$services;

   INST_ID NAME         GOAL         CLB_G
---------- -------------------- ------------
     1 PROD             NONE         LONG
     1 PROD_PRIM        NONE         LONG
     1 SYS$BACKGROUND   NONE         SHORT
     1 SYS$USERS        NONE         SHORT
     2 PROD             NONE         LONG
     2 PROD_PRIM        NONE         LONG
     2 SYS$BACKGROUND   NONE         SHORT
     2 SYS$USERS        NONE         SHORT

In order to change CLB goal as statements as per below

SQL> EXEC DBMS_SERVICE.MODIFY_SERVICE(service_name=>'PROD',goal=>2,clb_goal=>1);
SQL> EXEC DBMS_SERVICE.MODIFY_SERVICE(service_name=>'PROD_PRIM',goal=>2,clb_goal=>1);

 

GOODNESS Indicates how attractive a given instance is with respect to processing the workload that is presented to the service. A lower number is better. This number is internally computed based on the GOAL (LONG or SHORT) that is specified for the particular service.

SQL> select end_time,inst_id,service_name,cpupercall,goodness from gv$servicemetric where service_name like 'PROD%' order by end_time desc,inst_id;
 
END_TIME           INST_ID SERVICE_NAME         CPUPERCALL   GOODNESS
--------------- ---------- -------------------- ---------- ----------
03-JUL-12                2 PROD              2647.74074        800
03-JUL-12                1 PROD              406.694915       2300
03-JUL-12                1 PROD              1882.94935       2300
03-JUL-12                2 PROD              2024.03252        800

Testing:

In order to test these changes I will be using SwingBench tool which can simulate  Sales Order Entries transactions with a large number of sessions. SwingBench tool is open source and run on JAVA.

Swingbench will be setup as follows:

• Two desktop running swingbench will be used as load generators.

• A database called EBSTEST will be used. This is a 2 node RAC 11gR1 database.

• First desktop load generator will be runnig swingbench with 500 sessions, each session will reconnect after 5000 transaction have been completed. All this sessions will be forced to connect EBSTEST1 (instance 1) of the RAC cluster. Doing this will simulate the unbalanced situation, where 500+ sessions are connect to node1 and less sessions connected to node2. Also CPU utilization will be unbalanced.

• Second desktop load generator will be using the service_name EBSTEST which can conect either node of the cluster. Connecting to this server will consider LBA and CLB balancing algorithm. 500 session will be connected and each session will re-connect every 5000 transactions are completed.Next step is to un-assing the plan from the window

Note: First load generator can be seen here running 500 session and putting load in first node.

 Note: Load can be seen here first node LAVG is 31 where second node LAVG is 1.65.

 Note: sessions number is also unbalanced. 642 in instance 1 and 140 in instance 2.

Note: service load balancing is changed on the fly to consider CPU utilization. As you can see all new sessions are connected straight to instance 2 as you can see instance 1 still hold 644 sessions where instance 2 have gone from 140 to 338 sessions.

Note: as you can see node 2 LAVG has incresed from 1.6 to 4.39, confirming new  load balancing goal is now considering CPU utilization.

(HOWTO: Oracle DB) - Ebiz database running slow due to RESOURCE MANAGER feature is enabled

Ebiz database running slow due to RESOURCE MANAGER feature is enabled

RESOURCE MANAGER feature is enable in the database by default when you install 11g. This feature limit the amount of resources a process can use. This impact all payroll processes and other batch processes and cause session to wait on an event “resmgr:cpu quntum”.

This will hopefully remove the CPU limit that is causing “resmgr:cpu quntum” waits in your production database affecting concurrent programs mainly running during business hours.

Check database parameters

SQL> show parameter resource;

NAME TYPE VALUE

------------------------------------ ----------- ------------------------------

resource_limit boolean FALSE

resource_manager_cpu_allocation integer 8

resource_manager_plan string SCHEDULER[0x7AFFC]:DEFAULT_MAINTENANCE_PLAN

Disable resource manager from spfile

SQL> alter system set resource_manager_plan='' scope=both;

System altered.

Then, check resource manager windows

SQL> select window_name, resource_plan from dba_scheduler_windows;

WINDOW_NAME RESOURCE_PLAN

------------------------------ ------------------------------

WEEKNIGHT_WINDOW DEFAULT_MAINTENANCE_PLAN

WEEKEND_WINDOW DEFAULT_MAINTENANCE_PLAN

MONDAY_WINDOW DEFAULT_MAINTENANCE_PLAN

TUESDAY_WINDOW DEFAULT_MAINTENANCE_PLAN

WEDNESDAY_WINDOW DEFAULT_MAINTENANCE_PLAN

THURSDAY_WINDOW DEFAULT_MAINTENANCE_PLAN

FRIDAY_WINDOW DEFAULT_MAINTENANCE_PLAN

SATURDAY_WINDOW DEFAULT_MAINTENANCE_PLAN

SUNDAY_WINDOW DEFAULT_MAINTENANCE_PLAN

Next step is to un-assign the plan from the window

SQL> select 'execute dbms_scheduler.set_attribute('''||window_name||''',''RESOURCE_PLAN'','''');' from DBA_SCHEDULER_WINDOWS;

'EXECUTEDBMS_SCHEDULER.SET_ATTRIBUTE('''||WINDOW_NAME||''',''RESOURCE_PLAN'','''

--------------------------------------------------------------------------------

execute dbms_scheduler.set_attribute('WEEKNIGHT_WINDOW','RESOURCE_PLAN','');

execute dbms_scheduler.set_attribute('WEEKEND_WINDOW','RESOURCE_PLAN','');

execute dbms_scheduler.set_attribute('MONDAY_WINDOW','RESOURCE_PLAN','');

execute dbms_scheduler.set_attribute('TUESDAY_WINDOW','RESOURCE_PLAN','');

execute dbms_scheduler.set_attribute('WEDNESDAY_WINDOW','RESOURCE_PLAN','');

execute dbms_scheduler.set_attribute('THURSDAY_WINDOW','RESOURCE_PLAN','');

execute dbms_scheduler.set_attribute('FRIDAY_WINDOW','RESOURCE_PLAN','');

execute dbms_scheduler.set_attribute('SATURDAY_WINDOW','RESOURCE_PLAN','');

execute dbms_scheduler.set_attribute('SUNDAY_WINDOW','RESOURCE_PLAN','');

9 rows selected.

SQL> execute dbms_scheduler.set_attribute('WEEKNIGHT_WINDOW','RESOURCE_PLAN','');

PL/SQL procedure successfully completed.

...

...

SQL> commit;

Finally, disableauto tasks

SQL> BEGIN

DBMS_AUTO_TASK_ADMIN.DISABLE(

client_name => 'sql tuning advisor',

operation => NULL,

window_name => NULL);

END;

/

2 3 4 5 6 7

PL/SQL procedure successfully completed.

SQL> SQL> commit;

All Done.

Reference: High "Resmgr:Cpu Quantum" Wait Events In 11g Even When Resource Manager Is Disabled [ID 949033.1] 

HOWTO (Oracle DB) - Oradebug to trace session, hanganalyze and statedump

ORADEBUG Session Trace

The following steps are required to trace a user session with oradebug:

Obtain the SPID from v$process

SQL> select username, spid from v$process;

Start the debug session with the SPID of the process that needs traced

SQL> oradebug setospid 2280

SQL> oradebug unlimit

·         Select the appropriate trace level. There are four different options when specifying a tracing level.

·         Level 1 – provides “base set” tracing information. Bind variables are displayed as variables (:b1).

·         Level 4 – provides Level 1 data and the actual data values of bind variables.

·         Level 8 – provides Level 1 data and information on wait events when the elapsed time is greater than the CPU time.

·         Level 12 – combines levels 1, 4 and 8 tracing information. A Level 12 trace contains base set, bind variable values and wait events.

The oradebug command below will enable the maximum tracing possible:

SQL> oradebug event 10046 trace name context forever, level 12

Turn tracing off

SQL> oradebug event 10046 trace name context off

Obtain the trace file name. The oradebug facility provides an easy way to obtain the file name

SQL> oradebug tracefile_name

c:\oracle9i\admin\ORCL92\udump\mooracle_ora_2280.trc

Format the trace file with tkprof (as described in the earlier section on tkprof)

The result will be a trace file that contains more trace file information. Viewing wait events and bind variable values can be critical to diagnosing performance issues.

System state dump

In RAC system this need to be executed on all the instance

# Logon to sqlplus as sysdba

SQL> oradebug setmypid

SQL> oradebug unlimit

SQL> oradebug dump systemstate 10

... Wait at least 1 min

SQL> oradebug dump systemstate 10

... Wait at lease 1 min

SQL> oradebug dump systemstate 10

Hanganalyze dump

Syntax hangaanalyze for single instance

# Logon to sqlplus as sysdba

SQL> oradebug setmypid;

SQL> oradebug unlimit;

SQL> oradebug hanganalyze 3;

Syntax for hangaanalyze for RAC

# Logon to sqlplus as sysdba

SQL> oradebug setmypid;

SQL> oradebug unlimit;

SQL> oradebug setinst all

SQL> oradebug -g def hanganalyze 3

Connecting to SGA in case sqlplus does not work

$ sqlplus -prelim '/as sysdba

SQL> oradbug ...

HOWTO (ebiz) - Howto to restart OPP without having to restart full Concurrent Managers

This post describes the steps to restart OPP processes manually without having to take a full outage on concurrent managers.

Sometimes OPP does not re-start normally when DBA tries to re-start OPP from the Form "Concurrent Manager -> Administer" form.

Login to APPS and navigate to Concurrent Managers/Administer form

Query for %Output Post Processor% manager.

You will see Actual=4 and Target=4 processes

Click on Restart Button. Make sure Oputput Post Processor is selected.

Wait for at least 2 minutes.

You will see Target=0 and Actual=0

If Actual is still 4, then proceed with steps below to manual kill OPP processes.

Identify OPP process in the UNIX
Login to APPTIER server and identify the OPP processes.

[appl@papp1 log]$ ps -ef |grep FNDOPP | grep -v grep
appl  6947  7203  2 10:31 ?        00:01:21 /u01/oracle/PROD/apps/tech_st/10.1.3/appsutil/jdk/bin/java -DCLIENT_PROCESSID=6947 -server -Xmx384m -XX:NewRatio=2 -XX:+UseSerialGC -Doracle.apps.fnd.common.Pool.leak.mode=stderr:off -verbose:gc -mx1024m -Ddbcfile=/u01/oracle/PROD/inst/apps/PROD_papp1/appl/fnd/12.0.0/secure/PROD.dbc -Dcpid=486574 -Dconc_queue_id=6269 -Dqueue_appl_id=0 -Dlogfile=/u01/oracle/PROD/apps/logs/appl/conc/log/FNDOPP486574.txt -DLONG_RUNNING_JVM=true -DOVERRIDE_DBC=true -DFND_JDBC_BUFFER_MIN=1 -DFND_JDBC_BUFFER_MAX=2 oracle.apps.fnd.cp.gsf.GSMServiceController
appl  6948  7203  3 10:31 ?        00:01:30 /u01/oracle/PROD/apps/tech_st/10.1.3/appsutil/jdk/bin/java -DCLIENT_PROCESSID=6948 -server -Xmx384m -XX:NewRatio=2 -XX:+UseSerialGC -Doracle.apps.fnd.common.Pool.leak.mode=stderr:off -verbose:gc -mx1024m -Ddbcfile=/u01/oracle/PROD/inst/apps/PROD_papp1/appl/fnd/12.0.0/secure/PROD.dbc -Dcpid=486575 -Dconc_queue_id=6269 -Dqueue_appl_id=0 -Dlogfile=/u01/oracle/PROD/apps/logs/appl/conc/log/FNDOPP486575.txt -DLONG_RUNNING_JVM=true -DOVERRIDE_DBC=true -DFND_JDBC_BUFFER_MIN=1 -DFND_JDBC_BUFFER_MAX=2 oracle.apps.fnd.cp.gsf.GSMServiceController
appl  6949  7203  2 10:31 ?        00:01:06 /u01/oracle/PROD/apps/tech_st/10.1.3/appsutil/jdk/bin/java -DCLIENT_PROCESSID=6949 -server -Xmx384m -XX:NewRatio=2 -XX:+UseSerialGC -Doracle.apps.fnd.common.Pool.leak.mode=stderr:off -verbose:gc -mx1024m -Ddbcfile=/u01/oracle/PROD/inst/apps/PROD_papp1/appl/fnd/12.0.0/secure/PROD.dbc -Dcpid=486576 -Dconc_queue_id=6269 -Dqueue_appl_id=0 -Dlogfile=/u01/oracle/PROD/apps/logs/appl/conc/log/FNDOPP486576.txt -DLONG_RUNNING_JVM=true -DOVERRIDE_DBC=true -DFND_JDBC_BUFFER_MIN=1 -DFND_JDBC_BUFFER_MAX=2 oracle.apps.fnd.cp.gsf.GSMServiceController
appl  6953  7203  0 10:31 ?        00:00:27 /u01/oracle/PROD/apps/tech_st/10.1.3/appsutil/jdk/bin/java -DCLIENT_PROCESSID=6953 -server -Xmx384m -XX:NewRatio=2 -XX:+UseSerialGC -Doracle.apps.fnd.common.Pool.leak.mode=stderr:off -verbose:gc -mx1024m -Ddbcfile=/u01/oracle/PROD/inst/apps/PROD_papp1/appl/fnd/12.0.0/secure/PROD.dbc -Dcpid=486577 -Dconc_queue_id=6269 -Dqueue_appl_id=0 -Dlogfile=/u01/oracle/PROD/apps/logs/appl/conc/log/FNDOPP486577.txt -DLONG_RUNNING_JVM=true -DOVERRIDE_DBC=true -DFND_JDBC_BUFFER_MIN=1 -DFND_JDBC_BUFFER_MAX=2 oracle.apps.fnd.cp.gsf.GSMServiceController

Kill OPP processes

By Killing the processes you are forcing OPP process to stop. ICM will detect all OPP processes were found dead and Manually DBA has to Activate OPP manager from the Administer (concurrent managers) form.

$ kill -9 6953  6949  6948  6947

Tail the Main concurrent Manager log and confirm OPP processes have been found dead

$ cd $APPLCSF/log/PROD_1017.mgr

NOTE 1017 is the month and date of the last time concurrent managers started. Output is is a follows. You need to confirm all four processes are found dead.

                   Process monitor session started : 10-NOV-2011 10:29:13
Found dead process: spid=(7307), cpid=(485567), Service Instance=(6269)
Found dead process: spid=(7295), cpid=(485566), Service Instance=(6269)
Found dead process: spid=(7278), cpid=(485565), Service Instance=(6269)
Found dead process: spid=(7268), cpid=(485564), Service Instance=(6269)

Confirm Actual and target processes count is 0

Navigate to Concurrent Managers/Administer form and query for %Output Post Processor% manager.

You will see Actual=0 and Target=0 processes

Click on Activate Button. Make sure Oputput Post Processor is selected.

Wait for at least 2 minutes.

You will see Target=4 and Actual=4

Tail the Main concurrent Manager log and confirm OPP processes have started

$ cd $APPLCSF/log/PROD_1017.mgr

NOTE 1017 is the month and date of the last time concurrent managers started. Output is a follows. You need to confirm all four processes have started.

                 Process monitor session started : 10-NOV-2011 10:31:50
Starting FNDCPOPP Concurrent Manager               : 10-NOV-2011 10:31:51
Starting FNDCPOPP Concurrent Manager               : 10-NOV-2011 10:31:51
Starting FNDCPOPP Concurrent Manager               : 10-NOV-2011 10:31:51
Starting FNDCPOPP Concurrent Manager               : 10-NOV-2011 10:31:52

Check OPP log files are processing concurrent requests post actions correctly

[appl@papp1 log]$ $ cd $APPLCSF/log/FNDOPP*

check timestamp, normally last four processes

[appl@papp1 log]$ ls -lrt FNDOPP*
-rw-rw-r-- 1 appl dba 2836639 Nov  9 20:10 FNDOPP485566.txt
-rw-rw-r-- 1 appl dba 5704431 Nov  9 20:10 FNDOPP485565.txt
-rw-rw-r-- 1 appl dba 1316142 Nov 10 09:09 FNDOPP485567.txt
-rw-rw-r-- 1 appl dba 1016488 Nov 10 10:21 FNDOPP485564.txt
-rw-rw-r-- 1 appl dba   43403 Nov 10 11:59 FNDOPP486577.txt
-rw-rw-r-- 1 appl dba   47016 Nov 10 11:59 FNDOPP486576.txt
-rw-rw-r-- 1 appl dba   37590 Nov 10 11:59 FNDOPP486575.txt
-rw-rw-r-- 1 appl dba   42614 Nov 10 11:59 FNDOPP486574.txt

[appl@papp1 log]$ /sbin/fuser FNDOPP486574.txt FNDOPP486575.txt FNDOPP486576.txt FNDOPP486577.txt
FNDOPP486574.txt:     6947
FNDOPP486575.txt:     6948
FNDOPP486576.txt:     6949
FNDOPP486577.txt:     6953

Output file FNDOPP486574.txt

[11/10/11 10:34:43 AM] [486574:RT10109319] Executing post-processing actions for request 10109319.
[11/10/11 10:34:43 AM] [486574:RT10109319] Starting XML Publisher post-processing action.
[11/10/11 10:34:43 AM] [486574:RT10109319]
Template code: X142R
Template app:  XC
Language:      en
Territory:     null
Output type:   PDF
...
...
...
[11/10/11 10:34:44 AM] [486574:RT10109319] XML Publisher post-processing action complete.
[11/10/11 10:34:44 AM] [486574:RT10109319] Completed post-processing actions for request 10109319.

HOWTO (Oracle DB) - Monitor the progress and ETA of your RMAN Backup or Duplicate

Have you ever wondered how good or bad your RMAN backup or duplicate is performing?

This happened to me very often, backup and restore or duplicate operations are business critical activities. Normally, when a critical restore of your Production database is happening, everyone is looking over you shoulder asking "Is it done? What is the ETA?". These are questions that sometimes are hard to answer, specially if you have slow backup infrastructure or slow disk or Tapes.

The good thing is that Oracle provides all the information required to predict the ETA of an RMAN operation. All the information in there in the dictionary views. I have make my life easier by creating a shell script that gives you a full picture of your RMAN operation.

With this script you tell:

• When the operation will be finished.
• How much MBytes have been done.
• What is the current throughput/sec of the operation
• and mainly, give that information to your manager watching over your shoulder every second!!!

Output of the script:

$ . oraenv

$ bash rman_duplicate_monitor_eta.sh 

 
-------- RMAN: timing information - Tue May 15 22:05:47 EST 2012 ----------
 
OPNAME                              SOFAR  TOTALWORK   PCT_DONE    ELA_MIN    ETA_MIN START_TIME          FINISH_TIME
------------------------------ ---------- ---------- ---------- ---------- ---------- ------------------- -------------------
RMAN: incremental datafile bac        638        656      97.26          1          0 15-05-2012 22:02:30 15-05-2012 22:05:47
RMAN: aggregate input              737488     738214       99.9        5.2          0 15-05-2012 21:59:13 15-05-2012 22:05:47
RMAN: incremental datafile bac        640          0        100          1            15-05-2012 22:02:30
RMAN: aggregate output             519231          0        100        1.9            15-05-2012 22:02:30
 
'DUPLICATE/RESTORETHROUGHPUT MBYTES_SOFAR MBYTES_PER_SEC NAME
---------------------------- ------------ -------------- --------------------------
DUPLICATE/RESTORE THROUGHPUT         22.3            .07 physical write total bytes
 
'BACKUPTHROUGHPUT MBYTES_SOFAR MBYTES_PER_SEC NAME
----------------- ------------ -------------- --------------------------
BACKUP THROUGHPUT         5868          50.59 physical read total bytes

Shell Script code below:

#!/bin/bash
# $Id: ora_watch.sh 24 2010-03-16 00:18:42Z cricci $
#=========================================================================+
#  Copyright (c) 2009 Christian Daniel Ricci, Sydney, Australia           |
#                          All rights reserved.                           |
#=========================================================================+
# FILENAME
#       rman_duplicate_monitor_eta.sh.sh
#
# DESCRIPTION
#       Monitor RMAN operations
#
# HISTORY
#       CR - 16-Mar-2010 - Created
#
#=========================================================================*/

while [ 1 -eq 1 ]; do

ps -ef | grep rman | grep -v grep > /dev/null

[ $? -eq 1 ] && exit 0

echo "-------- RMAN: timing information - `date` ----------"

sqlplus -s '/as sysdba' <<EOF

col opname for a30

col message for a30 wra on

col units for a15

set lin 300 pages 1000 feed off echo off

alter session set nls_date_format='dd-mm-yyyy hh24:mi:ss';

select OPNAME,

--message,

SOFAR,TOTALWORK,round((SOFAR*100/decode(TOTALWORK,0,SOFAR,TOTALWORK)),2) PCT_DONE,

--UNITS,

round(elapsed_seconds/60,1) ela_min,round(time_remaining/60,1) eta_min,

START_TIME,round(time_remaining/60,1)/60/24+sysdate finish_time

 from v\$session_longops

where

OPNAME like 'RMAN%'

and SOFAR != TOTALWORK

order by 8;

SELECT

  'DUPLICATE/RESTORE THROUGHPUT',

  round(SUM(v.value/1024/1024),1) mbytes_sofar,

  round(SUM(v.value     /1024/1024)/nvl((SELECT MIN(elapsed_seconds)

  FROM v\$session_longops

  WHERE OPNAME LIKE 'RMAN: aggregate input'

  AND SOFAR            != TOTALWORK

  AND elapsed_seconds IS NOT NULL

  ),SUM(v.value     /1024/1024)),2) mbytes_per_sec,

  n.name

FROM gv\$sesstat v,

  v\$statname n,

  gv\$session s

WHERE v.statistic#=n.statistic#

AND n.name = 'physical write total bytes'

AND v.sid = s.sid

AND v.inst_id=s.inst_id

AND s.program like 'rman@%'

GROUP BY 'DUPLICATE/RESTORE THROUGHPUT',n.name;

SELECT

  'BACKUP THROUGHPUT',

  round(SUM(v.value/1024/1024),1) mbytes_sofar,

  round(SUM(v.value     /1024/1024)/nvl((SELECT MIN(elapsed_seconds)

  FROM v\$session_longops

  WHERE OPNAME LIKE 'RMAN: aggregate output'

  AND SOFAR            != TOTALWORK

  AND elapsed_seconds IS NOT NULL

  ),SUM(v.value     /1024/1024)),2) mbytes_per_sec,

  n.name

FROM gv\$sesstat v,

  v\$statname n,

  gv\$session s

WHERE v.statistic#=n.statistic#

AND n.name = 'physical read total bytes'

AND v.sid = s.sid

AND v.inst_id=s.inst_id

AND s.program like 'rman@%'

GROUP BY 'BACKUP THROUGHPUT',n.name;

exit

EOF

sleep 60

done

  

NON-TECHIE (Your health) - Sweet Poison

NON-TECHIE (Your health) - Sweet Poison

This is an email it was sent to me from one of my friends, the subject worries me, since my friend works in the health are.

This is what email says... in quotes

In October of 2001, my sister started getting very sick.
She had stomach spasms and she was having a hard time getting around.
Walking was a major chore.
It took everything she had just to get out of bed; she was in so much pain.


By March 2002, she had undergone several tissue and muscle biopsies
and was on 24 various prescription medications.
The doctors could not determine what was wrong with her.
She was in so much pain, and so sick she just knew she was dying.


She put her house, bank accounts, life insurance, etc., in her oldest
daughter's name, and made sure that her younger children were to be
taken care of.


She also wanted her last hooray, so she planned a trip to Florida
(basically in a wheelchair) for March 22nd.


On March 19 I called her to ask how her most recent tests went, and
she said they didn't find anything on the test, but they believe she
had MS.
I recalled an article a friend of mine e-mailed to me and I asked my
sister if she drank diet soda?
She told me that she did.
As a matter of fact, she was getting ready to crack one open that moment.


I told her not to open it, and to stop drinking the diet soda!
I e-mailed her article my friend, a lawyer, had sent.
My sister called me within 32 hours after our phone conversation and
told me she had stopped drinking the diet soda AND she could walk!
The muscle spasms went away. She said she didn't feel 100% but, she
sure felt a lot better.


She told me she was going to her doctor with this article and would
call me when she got home.


Well, she called me, and said her doctor was amazed!
He is going to call all of his MS patients to find out if they
consumed artificial sweeteners of any kind.
In a nutshell, she was being poisoned by the Aspartame in the diet
soda...and literally dying a slow and miserable death.


When she got to Florida March 22, all she had to take was one pill,
and that was a pill for the Aspartame poisoning!
She is well on her way to a complete recovery.
And she is walking!
No wheelchair!


This article saved her life.
If it says 'SUGAR FREE' on the label;
DO NOT EVEN THINK ABOUT IT!
I have spent several days lecturing at the WORLD ENVIRONMENTAL
CONFERENCE on 'ASPARTAME,' marketed as 'Nutra Sweet,'
'Equal,' and
'Spoonful.'


In the keynote address by the EPA, it was announced that in the United
States in 2001 there is an epidemic of multiple sclerosis and systemic
lupus.
It was difficult to determine exactly what toxin was causing this to
be rampant.
I stood up and said that I was there to lecture on exactly that subject.
I will explain why Aspartame is so dangerous:
When the temperature of this sweetener exceeds 86 degrees F, the wood
alcohol in ASPARTAME converts to formaldehyde and then to formic acid,
which in turn causes metabolic acidosis.. Formic acid is the poison
found in the sting of fire ants.
The methanol toxicity mimics, among other conditions, multiple
sclerosis and systemic lupus.


Many people were being diagnosed in error.
Although multiple sclerosis is not a death sentence, Methanol toxicity is!


Systemic lupus has become almost as rampant as multiple sclerosis,
especially with Diet Coke and Diet Pepsi drinkers.


The victim usually does not know that the Aspartame is the culprit.
He or she continues its use; irritating the lupus to such a degree
that it may become a life-threatening condition.
We have seen patients with systemic lupus become asymptotic, once
taken off diet sodas.


In cases of those diagnosed with Multiple Sclerosis, most of the
symptoms disappear. We've seen many cases where vision loss returned
and hearing loss improved markedly.


This also applies to cases of tinnitus and fibromyalgia.
During a lecture, I said,
'If you are using ASPARTAME (Nutra Sweet, Equal, Spoonful, etc) and
you suffer from fibromyalgia symptoms, spasms, shooting, pains,
numbness in your legs,
Cramps,
Vertigo,
Dizziness,
Headaches,
Tinnitus,
Joint pain,
Unexplainable depression, anxiety attacks, slurred speech, blurred
vision, or memory loss you probably have ASPARTAME poisoning!'
People were jumping up during the lecture saying,
'I have some of these symptoms..
Is it reversible?'


Yes!
Yes!
Yes!
STOP drinking diet sodas and be alert for Aspartame on food labels!
Many products are fortified with it!
This is a serious problem.
Dr. Espart (one of my speakers) remarked that so many people seem to
be symptomatic for MS and during his recent visit to a hospice; a
nurse stated that six of her friends, who were heavy Diet Coke
addicts, had all been diagnosed with MS. This is beyond coincidence!


Diet soda is NOT a diet product! It is a chemically altered, multiple
SODIUM (salt) and ASPARTAME containing product that actually makes you
crave carbohydrates.


It is far more likely to make you GAIN weight!


These products also contain formaldehyde, which stores in the fat
cells, particularly in the hips and thighs.
Formaldehyde is an absolute toxin and is used primarily to preserve
'tissue specimens.'


Many products we use every day contain this chemical but we SHOULD NOT
store it IN our body!


Dr. H. J. Roberts stated in his lectures that once free of the 'diet
products' and with no significant increase in exercise; his patients
lost an average of 19 pounds over a trial period.


Aspartame is especially dangerous for diabetics.
We found that some physicians, who believed that they had a patient
with retinopathy, in fact, had symptoms caused by Aspartame.
The Aspartame drives the blood sugar out of control.
Thus diabetics may suffer acute memory loss due to the fact that
aspartic acid and phenylalanine are NEUROTOXIC when taken without the
other amino acids necessary for a good balance.


Treating diabetes is all about BALANCE.
Especially with diabetics, the Aspartame passes the blood/brain
barrier and it then deteriorates the neurons of the brain; causing
various levels of brain damage,
Seizures,
Depression,
Manic depression,
Panic attacks,
Uncontrollable anger and rage.


Consumption of Aspartame causes these same symptoms in non-diabetics as well.
Documentation and observation also reveal that thousands of children
diagnosed with ADD and ADHD have had complete turnarounds in their
behavior when these chemicals have been removed from their diet.


So called 'behavior modification prescription drugs' (Ritalin and
others) are no longer needed.
Truth be told, they were never NEEDED in the first place!


Most of these children were being 'poisoned' on a daily basis with the
very foods that were 'better for them than sugar.'


It is also suspected that the Aspartame in thousands of pallets ofdiet
Coke and diet Pepsiconsumed by men and women fighting in the Gulf War,
may be partially to blame for the well-known Gulf War Syndrome.


Dr. Roberts warns that it can cause birth defects, i.e. mental
retardation, if taken at the time of conception and during early
pregnancy.
Children are especially at risk for neurological disorders and should
NEVER be given artificial sweeteners.


There are many different case histories to relate of children
suffering grand mal seizures and other neurological disturbances
talking about a plague of neurological diseases directly caused by the
use of this deadly poison.'


Herein lies the problem:
There were Congressional Hearings when Aspartame was included in 100
different products and strong objection was made concerning its use.
Since this initial hearing, there have been two subsequent hearings,
and still nothing has been done.
The drug and chemical lobbies have very deep pockets.


Sadly, MONSANTO'S patent on Aspartame has EXPIRED!
There are now over 5,000 products on the market that contain this
deadly chemical and there will be thousands more introduced.
Everybody wants a 'piece of the Aspartame pie.'


I assure you that MONSANTO, the creator of Aspartame, knows how deadly it is.


And isn't it ironic that MONSANTO funds, among others, the American
Diabetes   Association, the American Dietetic Association and the
Conference of the American College of Physicians?


This has been recently exposed in the New York Times.
These [organizations] cannot criticize any additives or convey their
link to MONSANTO because they take money from the food industry and
are required to endorse their products.


Senator Howard Metzenbaum wrote and presented a bill that would
require label warnings on products containing Aspartame, especially
regarding pregnant women, children and infants.


The bill would also institute independent studies on the known dangers
and the problems existing in the general population regarding
seizures, changes in brain chemistry, neurological changes and
behavioral symptoms.


The bill was killed.


It is known that the powerful drug and chemical lobbies are
responsible for this, letting loose the hounds of disease and death on
an unsuspecting and uninformed public. Well, you're informed now!


YOU HAVE A RIGHT TO KNOW!