Sequences in 10G RAC

Sequences in 10G RAC

Recently we migrated our application to Oracle 10g RAC (3 Instance) and ended up with sequences created as out of order at each Instance.

Here is the snippet of code to demonstrate that.

rajesh@ORA10GR2-01> SELECT INST_ID,
  2    INSTANCE_NUMBER,
  3    INSTANCE_NAME
  4  FROM GV$INSTANCE
  5  ORDER BY INST_ID;

   INST_ID INSTANCE_NUMBER INSTANCE_NAME
---------- --------------- ----------------
         1               1 ORA10GR2-01
         2               2 ORA10GR2-02
         3               3 ORA10GR2-03


 rajesh@ORA10GR2-01> create sequence s CACHE 30;

Sequence created.

Elapsed: 00:00:00.07
rajesh@ORA10GR2-01> select SEQUENCE_NAME,ORDER_FLAG,CACHE_SIZE
  2  from user_sequences
  3  where sequence_name ='S'
  4  /

SEQUENCE_NAME                  O CACHE_SIZE
------------------------------ - ----------
S                              N         30

Elapsed: 00:00:00.17
rajesh@ORA10GR2-01>
rajesh@ORA10GR2-01>
rajesh@ORA10GR2-01> select s.nextval from dual;

   NEXTVAL
----------
         1

Elapsed: 00:00:00.14
rajesh@ORA10GR2-01>
rajesh@ORA10GR2-01> select s.nextval from dual;

   NEXTVAL
----------
         2

Elapsed: 00:00:00.25
rajesh@ORA10GR2-01>

When connecting back to Instance02 I found the sequence values different.

rajesh@ORA10GR2-01> connect rajesh/oracle@ORA10GR2-02
Connected.
rajesh@ORA10GR2-02> select s.nextval from dual;

   NEXTVAL
----------
        31

Elapsed: 00:00:00.48
rajesh@ORA10GR2-02>
rajesh@ORA10GR2-02> select s.nextval from dual;

   NEXTVAL
----------
        32

Elapsed: 00:00:00.15
rajesh@ORA10GR2-02>

When connecting back to Instance03 I found the sequence values different again.

rajesh@ORA10GR2-02> connect rajesh/oracle@ORA10GR2-03
Connected.
rajesh@ORA10GR2-03> select s.nextval from dual;

   NEXTVAL
----------
        61

Elapsed: 00:00:00.21
rajesh@ORA10GR2-03> select s.nextval from dual;

   NEXTVAL
----------
        62

Elapsed: 00:00:00.14
rajesh@ORA10GR2-03>
The reason for this out of order sequence at each instance is the sequence cache is in the shared pool, therefore sessions on the same node can share the cached entry, but sessions on different nodes cannot.

The solution to solve this sequence running out of order at each instance is to create sequence using ORDER option. which will guarantee the order in which numbers are generated.

rajesh@ORA10GR2-01> create sequence s CACHE 30 ORDER;

Sequence created.

Elapsed: 00:00:00.09
rajesh@ORA10GR2-01>
rajesh@ORA10GR2-01> select SEQUENCE_NAME,ORDER_FLAG,CACHE_SIZE
  2  from user_sequences
  3  where sequence_name ='S'
  4  /

SEQUENCE_NAME                  O CACHE_SIZE
------------------------------ - ----------
S                              Y         30

Elapsed: 00:00:00.10
rajesh@ORA10GR2-01>
rajesh@ORA10GR2-01> select s.nextval from dual;

   NEXTVAL
----------
         1

Elapsed: 00:00:00.17
rajesh@ORA10GR2-01> select s.nextval from dual;

   NEXTVAL
----------
         2

Elapsed: 00:00:00.34
rajesh@ORA10GR2-01>
Connecting back to Instance02.

rajesh@ORA10GR2-01> connect rajesh/oracle@ORA10GR2-02

Connected.
rajesh@ORA10GR2-02> select s.nextval from dual;

   NEXTVAL
----------
         3

Elapsed: 00:00:00.17
rajesh@ORA10GR2-02>
rajesh@ORA10GR2-02> select s.nextval from dual;

   NEXTVAL
----------
         4

Elapsed: 00:00:00.17
rajesh@ORA10GR2-02>
Connecting back to Instance03.

rajesh@ORA10GR2-02> connect rajesh/oracle@ORA10GR2-03
Connected.
rajesh@ORA10GR2-03> select s.nextval from dual;

   NEXTVAL
----------
         5

Elapsed: 00:00:00.14
rajesh@ORA10GR2-03>
rajesh@ORA10GR2-03> select s.nextval from dual;

   NEXTVAL
----------
         6

Elapsed: 00:00:00.17
rajesh@ORA10GR2-03>

We changed the sequence to "ordered" Now selecting on either node gets the next number as expected.

However having the sequence with CACHE and ORDER would lead to some performance implications due to cluster synchronization. In Oracle 10g RAC, if you specify the "ordered" clause for a sequence, then a global lock is allocated by the node when you access the sequence. This lock acquisition happens only at the first sequence access for the node, and subsequent uses of the sequence do not wait on this lock.


rajesh@ORA10GR2-01> create sequence s order;

Sequence created.

Elapsed: 00:00:00.01
rajesh@ORA10GR2-01>
rajesh@ORA10GR2-01> select SEQUENCE_NAME,ORDER_FLAG,CACHE_SIZE
  2  from user_sequences
  3  where sequence_name ='S';

SEQUENCE_NAME                  O CACHE_SIZE
------------------------------ - ----------
S                              Y         20

Elapsed: 00:00:00.01
rajesh@ORA10GR2-01>
rajesh@ORA10GR2-01>
Connected.
rajesh@ORA10GR2-01>

I was running the below Pl/SQL block evenly across all the RAC Instance.

rajesh@ORA10GR2-01>  declare
  2             k number;
  3             l_start number;
  4             l_cpu_time number;
  5     begin
  6             l_start := dbms_utility.get_time;
  7             l_cpu_time := dbms_utility.get_cpu_time;
  8     for i in 1..500000
  9     loop
 10                     select s.nextval
 11                     into k
 12                     from dual;
 13     end loop;
 14
 15             dbms_output.put_line (' Total Execution time = '|| (dbms_utility.get_time - l_start) );
 16             dbms_output.put_line (' Total CPU time = '|| (dbms_utility.get_cpu_time - l_cpu_time) );
 17     end;
 18  /

 Total Execution time = 82110
 Total CPU time = 7057

PL/SQL procedure successfully completed.

Elapsed: 00:13:40.99
rajesh@ORA10GR2-01>

And the AWR on Instance 1 is.

   SQL Id      Time (ms)
------------- ----------
9x2hyv1prpts9    802,038
Module: SQL*Plus
OUTPUT
--------------------------------------------------------------------------------
 declare k number; l_start number; l_cpu_time number; begin l_start :=
dbms_utility.get_time; l_cpu_time := dbms_utility.get_cpu_time; for i in 1..5
00000 loop select s.nextval into k from dual; end loop; dbms_outpu
t.put_line (' Total Execution time = '|| (dbms_utility.get_time - l_start) );
          -------------------------------------------------------------
SQL ID: 9x2hyv1prpts9              DB/Inst: IRDST01/irdst011  Snaps: 3518-3519
-> 1st Capture and Last Capture Snap IDs
   refer to Snapshot IDs witin the snapshot range
-> declare k number; l_start number; l_cpu_time number; begin l...
OUTPUT
--------------------------------------------------------------------------------
    Plan Hash           Total Elapsed                 1st Capture   Last Capture
#   Value                    Time(ms)    Executions       Snap ID        Snap ID
--- ---------------- ---------------- ------------- ------------- --------------
1   0                         802,038             1          3519           3519
          -------------------------------------------------------------
Plan 1(PHV: 0)
--------------
OUTPUT
--------------------------------------------------------------------------------
Plan Statistics                    DB/Inst: IRDST01/irdst011  Snaps: 3518-3519
-> % Total DB Time is the Elapsed Time of the SQL statement divided
   into the Total Database Time multiplied by 100
Stat Name                                Statement   Per Execution % Snap
---------------------------------------- ---------- -------------- -------
Elapsed Time (ms)                           802,038      802,037.9    30.0
CPU Time (ms)                                70,597       70,597.3    93.7
Executions                                        1            N/A     N/A
Buffer Gets                                 149,976      149,976.0    56.6
Disk Reads                                        3            3.0     7.7
OUTPUT
--------------------------------------------------------------------------------
Parse Calls                                       1            1.0     0.0
Rows                                              1            1.0     N/A
User I/O Wait Time (ms)                          45            N/A     N/A
Cluster Wait Time (ms)                      186,959            N/A     N/AApplication Wait Time (ms)                        0            N/A     N/A
Concurrency Wait Time (ms)                   13,813            N/A     N/AInvalidations                                     0            N/A     N/A
Version Count                                     1            N/A     N/A
Sharable Mem(KB)                                 20            N/A     N/A
          -------------------------------------------------------------

And the AWR on Instance 2 is.

                Elapsed
   SQL Id      Time (ms)
------------- ----------
9x2hyv1prpts9    818,718
Module: SQL*Plus
OUTPUT
--------------------------------------------------------------------------------
 declare k number; l_start number; l_cpu_time number; begin l_start :=
dbms_utility.get_time; l_cpu_time := dbms_utility.get_cpu_time; for i in 1..5
00000 loop select s.nextval into k from dual; end loop; dbms_outpu
t.put_line (' Total Execution time = '|| (dbms_utility.get_time - l_start) );
          -------------------------------------------------------------
SQL ID: 9x2hyv1prpts9              DB/Inst: IRDST01/irdst012  Snaps: 3518-3519
-> 1st Capture and Last Capture Snap IDs
   refer to Snapshot IDs witin the snapshot range
-> declare k number; l_start number; l_cpu_time number; begin l...
OUTPUT
--------------------------------------------------------------------------------
    Plan Hash           Total Elapsed                 1st Capture   Last Capture
#   Value                    Time(ms)    Executions       Snap ID        Snap ID
--- ---------------- ---------------- ------------- ------------- --------------
1   0                         818,718             1          3519           3519
          -------------------------------------------------------------
Plan 1(PHV: 0)
--------------
OUTPUT
--------------------------------------------------------------------------------
Plan Statistics                    DB/Inst: IRDST01/irdst012  Snaps: 3518-3519
-> % Total DB Time is the Elapsed Time of the SQL statement divided
   into the Total Database Time multiplied by 100
Stat Name                                Statement   Per Execution % Snap
---------------------------------------- ---------- -------------- -------
Elapsed Time (ms)                           818,718      818,718.5    31.7
CPU Time (ms)                                72,017       72,017.1    97.5
Executions                                        1            N/A     N/A
Buffer Gets                                 149,898      149,898.0    81.6
Disk Reads                                        0            0.0     0.0
OUTPUT
--------------------------------------------------------------------------------
Parse Calls                                       1            1.0     0.0
Rows                                              1            1.0     N/A
User I/O Wait Time (ms)                           0            N/A     N/A
Cluster Wait Time (ms)                      135,191            N/A     N/A
Application Wait Time (ms)                        0            N/A     N/A
Concurrency Wait Time (ms)                   14,768            N/A     N/A
Invalidations                                     0            N/A     N/A
Version Count                                     1            N/A     N/A
Sharable Mem(KB)                                 20            N/A     N/A
          -------------------------------------------------------------

And the AWR on Instance 3 is.

                Elapsed
   SQL Id      Time (ms)
------------- ----------
9x2hyv1prpts9    816,826
Module: SQL*Plus
OUTPUT
--------------------------------------------------------------------------------
 declare k number; l_start number; l_cpu_time number; begin l_start :=
dbms_utility.get_time; l_cpu_time := dbms_utility.get_cpu_time; for i in 1..5
00000 loop select s.nextval into k from dual; end loop; dbms_outpu
t.put_line (' Total Execution time = '|| (dbms_utility.get_time - l_start) );
          -------------------------------------------------------------
SQL ID: 9x2hyv1prpts9              DB/Inst: IRDST01/irdst013  Snaps: 3518-3519
-> 1st Capture and Last Capture Snap IDs
   refer to Snapshot IDs witin the snapshot range
-> declare k number; l_start number; l_cpu_time number; begin l...
OUTPUT
--------------------------------------------------------------------------------
    Plan Hash           Total Elapsed                 1st Capture   Last Capture
#   Value                    Time(ms)    Executions       Snap ID        Snap ID
--- ---------------- ---------------- ------------- ------------- --------------
1   0                         816,826             1          3519           3519
          -------------------------------------------------------------
Plan 1(PHV: 0)
--------------
OUTPUT
--------------------------------------------------------------------------------
Plan Statistics                    DB/Inst: IRDST01/irdst013  Snaps: 3518-3519
-> % Total DB Time is the Elapsed Time of the SQL statement divided
   into the Total Database Time multiplied by 100
Stat Name                                Statement   Per Execution % Snap
---------------------------------------- ---------- -------------- -------
Elapsed Time (ms)                           816,826      816,826.4    97.8
CPU Time (ms)                                68,859       68,858.5    90.9
Executions                                        1            N/A     N/A
Buffer Gets                                 133,660      133,660.0    28.3
Disk Reads                                        0            0.0     0.0
OUTPUT
--------------------------------------------------------------------------------
Parse Calls                                       1            1.0     0.0
Rows                                              1            1.0     N/A
User I/O Wait Time (ms)                           0            N/A     N/A
Cluster Wait Time (ms)                       25,148            N/A     N/A
Application Wait Time (ms)                        0            N/A     N/A
Concurrency Wait Time (ms)                    8,942            N/A     N/A
Invalidations                                     0            N/A     N/A
Version Count                                     1            N/A     N/A
Sharable Mem(KB)                                 20            N/A     N/A
          -------------------------------------------------------------

Due to the less CACHE sequence value and ORDER attribute and since the same sequence is shared across all Instance we ended up with Clustered & Concurrency waits.

Now repeting the same test with Increased CACHE size.

rajesh@ORA10GR2-01> create sequence s CACHE 100000 order;

Sequence created.

Elapsed: 00:00:00.03
rajesh@ORA10GR2-01>
rajesh@ORA10GR2-01> select SEQUENCE_NAME,ORDER_FLAG,CACHE_SIZE
  2  from user_sequences
  3  where sequence_name ='S';

SEQUENCE_NAME                  O CACHE_SIZE
------------------------------ - ----------
S                              Y     100000

Elapsed: 00:00:00.01 

rajesh@ORA10GR2-01>  declare
  2             k number;
  3             l_start number;
  4             l_cpu_time number;
  5     begin
  6             l_start := dbms_utility.get_time;
  7             l_cpu_time := dbms_utility.get_cpu_time;
  8     for i in 1..500000
  9     loop
 10                     select s.nextval
 11                     into k
 12                     from dual;
 13     end loop;
 14
 15             dbms_output.put_line (' Total Execution time = '|| (dbms_utility.get_time - l_start) );
 16             dbms_output.put_line (' Total CPU time = '|| (dbms_utility.get_cpu_time - l_cpu_time) );
 17     end;
 18  /

 Total Execution time = 32396
 Total CPU time = 3418

PL/SQL procedure successfully completed.

Elapsed: 00:05:23.96
rajesh@ORA10GR2-01>

AWR at Instance-01 shows

                Elapsed
   SQL Id      Time (ms)
------------- ----------
9x2hyv1prpts9    316,430
Module: SQL*Plus
OUTPUT
--------------------------------------------------------------------------------
 declare k number; l_start number; l_cpu_time number; begin l_start :=
dbms_utility.get_time; l_cpu_time := dbms_utility.get_cpu_time; for i in 1..5
00000 loop select s.nextval into k from dual; end loop; dbms_outpu
t.put_line (' Total Execution time = '|| (dbms_utility.get_time - l_start) );
          -------------------------------------------------------------
SQL ID: 9x2hyv1prpts9              DB/Inst: IRDST01/irdst011  Snaps: 3520-3521
-> 1st Capture and Last Capture Snap IDs
   refer to Snapshot IDs witin the snapshot range
-> declare k number; l_start number; l_cpu_time number; begin l...
OUTPUT
--------------------------------------------------------------------------------
    Plan Hash           Total Elapsed                 1st Capture   Last Capture
#   Value                    Time(ms)    Executions       Snap ID        Snap ID
--- ---------------- ---------------- ------------- ------------- --------------
1   0                         316,430             1          3521           3521
          -------------------------------------------------------------
Plan 1(PHV: 0)
--------------
OUTPUT
--------------------------------------------------------------------------------
Plan Statistics                    DB/Inst: IRDST01/irdst011  Snaps: 3520-3521
-> % Total DB Time is the Elapsed Time of the SQL statement divided
   into the Total Database Time multiplied by 100
Stat Name                                Statement   Per Execution % Snap
---------------------------------------- ---------- -------------- -------
Elapsed Time (ms)                           316,430      316,429.6    91.0
CPU Time (ms)                                34,199       34,198.8    87.9
Executions                                        1            N/A     N/A
Buffer Gets                                     118          118.0     0.9
Disk Reads                                        0            0.0     0.0
OUTPUT
--------------------------------------------------------------------------------
Parse Calls                                       1            1.0     0.1
Rows                                              1            1.0     N/A
User I/O Wait Time (ms)                           0            N/A     N/A
Cluster Wait Time (ms)                            4            N/A     N/AApplication Wait Time (ms)                        0            N/A     N/A
Concurrency Wait Time (ms)                        8            N/A     N/AInvalidations                                     0            N/A     N/A
Version Count                                     1            N/A     N/A
Sharable Mem(KB)                                 20            N/A     N/A
          -------------------------------------------------------------

AWR at Instance-02 shows

--------------------------------------------------------------------------------
 declare k number; l_start number; l_cpu_time number; begin l_start :=
dbms_utility.get_time; l_cpu_time := dbms_utility.get_cpu_time; for i in 1..5
00000 loop select s.nextval into k from dual; end loop; dbms_outpu
t.put_line (' Total Execution time = '|| (dbms_utility.get_time - l_start) );
          -------------------------------------------------------------
SQL ID: 9x2hyv1prpts9              DB/Inst: IRDST01/irdst012  Snaps: 3520-3521
-> 1st Capture and Last Capture Snap IDs
   refer to Snapshot IDs witin the snapshot range
-> declare k number; l_start number; l_cpu_time number; begin l...
OUTPUT
--------------------------------------------------------------------------------
    Plan Hash           Total Elapsed                 1st Capture   Last Capture
#   Value                    Time(ms)    Executions       Snap ID        Snap ID
--- ---------------- ---------------- ------------- ------------- --------------
1   0                         320,198             1          3521           3521
          -------------------------------------------------------------
Plan 1(PHV: 0)
--------------
OUTPUT
--------------------------------------------------------------------------------
Plan Statistics                    DB/Inst: IRDST01/irdst012  Snaps: 3520-3521
-> % Total DB Time is the Elapsed Time of the SQL statement divided
   into the Total Database Time multiplied by 100
Stat Name                                Statement   Per Execution % Snap
---------------------------------------- ---------- -------------- -------
Elapsed Time (ms)                           320,198      320,198.1    96.7
CPU Time (ms)                                36,561       36,561.4    92.0
Executions                                        1            N/A     N/A
Buffer Gets                                      59           59.0     0.7
Disk Reads                                        0            0.0     0.0
OUTPUT
--------------------------------------------------------------------------------
Parse Calls                                       1            1.0     0.1
Rows                                              1            1.0     N/A
User I/O Wait Time (ms)                           0            N/A     N/A
Cluster Wait Time (ms)                           12            N/A     N/A
Application Wait Time (ms)                        0            N/A     N/A
Concurrency Wait Time (ms)                        9            N/A     N/AInvalidations                                     0            N/A     N/A
Version Count                                     1            N/A     N/A
Sharable Mem(KB)                                 20            N/A     N/A
          -------------------------------------------------------------

AWR at Instance-03 shows

--------------------------------------------------------------------------------
 declare k number; l_start number; l_cpu_time number; begin l_start :=
dbms_utility.get_time; l_cpu_time := dbms_utility.get_cpu_time; for i in 1..5
00000 loop select s.nextval into k from dual; end loop; dbms_outpu
t.put_line (' Total Execution time = '|| (dbms_utility.get_time - l_start) );
          -------------------------------------------------------------
SQL ID: 9x2hyv1prpts9              DB/Inst: IRDST01/irdst013  Snaps: 3520-3521
-> 1st Capture and Last Capture Snap IDs
   refer to Snapshot IDs witin the snapshot range
-> declare k number; l_start number; l_cpu_time number; begin l...
OUTPUT
--------------------------------------------------------------------------------
    Plan Hash           Total Elapsed                 1st Capture   Last Capture
#   Value                    Time(ms)    Executions       Snap ID        Snap ID
--- ---------------- ---------------- ------------- ------------- --------------
1   0                         318,115             1          3521           3521
          -------------------------------------------------------------
Plan 1(PHV: 0)
--------------
OUTPUT
--------------------------------------------------------------------------------
Plan Statistics                    DB/Inst: IRDST01/irdst013  Snaps: 3520-3521
-> % Total DB Time is the Elapsed Time of the SQL statement divided
   into the Total Database Time multiplied by 100
Stat Name                                Statement   Per Execution % Snap
---------------------------------------- ---------- -------------- -------
Elapsed Time (ms)                           318,115      318,115.2    92.6
CPU Time (ms)                                31,745       31,745.2    83.8
Executions                                        1            N/A     N/A
Buffer Gets                                      47           47.0     0.0
Disk Reads                                        0            0.0     0.0
OUTPUT
--------------------------------------------------------------------------------
Parse Calls                                       1            1.0     0.1
Rows                                              1            1.0     N/A
User I/O Wait Time (ms)                           0            N/A     N/A
Cluster Wait Time (ms)                           11            N/A     N/A
Application Wait Time (ms)                        0            N/A     N/A
Concurrency Wait Time (ms)                        5            N/A     N/AInvalidations                                     0            N/A     N/A
Version Count                                     1            N/A     N/A
Sharable Mem(KB)                                 20            N/A     N/A
          -------------------------------------------------------------

Now the Cluster Wait Time & Concurrency Wait Time got reduced due to Increased CACHE size. which leads to less synchronization across RAC Instances.

Java Messaging Service - Consumer / Subscriber

package oc.test.olivia;


import javax.jms.*;

import org.apache.activemq.ActiveMQConnection;
import org.apache.activemq.ActiveMQConnectionFactory;

/*
 * @author Olivia Cassandrae
 */
public class Consumer {

    public static void main(String[] args) throws JMSException {
        ConnectionFactory connectionFactory
            = new ActiveMQConnectionFactory(url);
        Connection connection = connectionFactory.createConnection();
        connection.start();
        Session session = connection.createSession(false,
            Session.AUTO_ACKNOWLEDGE);
        Destination destination = session.createQueue(subject);
        MessageConsumer consumer = session.createConsumer(destination);
        Message message = consumer.receive();
        if (message instanceof TextMessage) {
            TextMessage textMessage = (TextMessage) message;
            System.out.println("Received message '"
                + textMessage.getText() + "'");
        }
        connection.close();
    }
   
    // URL of the JMS server
    private static String url = ActiveMQConnection.DEFAULT_BROKER_URL;
    private static String subject = "GROUNDOPS.PSCS.IROPS.BRIDGE.QUEUE";
}

Url: http://localhost:8161/admin

Java Messaging Service - Producer / Publisher

Exception:


ConnectionFactory is ::: org.apache.activemq.ActiveMQConnectionFactory
javax.jms.JMSException: Could not connect to broker URL: tcp://PDCHCL-TRAX9:61616. Reason: java.net.ConnectException: Connection refused: connect
at org.apache.activemq.util.JMSExceptionSupport.create(JMSExceptionSupport.java:35)
at org.apache.activemq.ActiveMQConnectionFactory.createActiveMQConnection(ActiveMQConnectionFactory.java:278)
at org.apache.activemq.ActiveMQConnectionFactory.createActiveMQConnection(ActiveMQConnectionFactory.java:222)
at org.apache.activemq.ActiveMQConnectionFactory.createQueueConnection(ActiveMQConnectionFactory.java:185)
at com.test.IROPSQueueApacheMQProducer.main(IROPSQueueApacheMQProducer.java:58)
Caused by: java.net.ConnectException: Connection refused: connect
at java.net.PlainSocketImpl.socketConnect(Native Method)
at java.net.PlainSocketImpl.doConnect(Unknown Source)
at java.net.PlainSocketImpl.connectToAddress(Unknown Source)
at java.net.PlainSocketImpl.connect(Unknown Source)
at java.net.SocksSocketImpl.connect(Unknown Source)
at java.net.Socket.connect(Unknown Source)
at org.apache.activemq.transport.tcp.TcpTransport.connect(TcpTransport.java:414)
at org.apache.activemq.transport.tcp.TcpTransport.doStart(TcpTransport.java:380)
at org.apache.activemq.util.ServiceSupport.start(ServiceSupport.java:50)
at org.apache.activemq.transport.TransportFilter.start(TransportFilter.java:57)
at org.apache.activemq.transport.TransportFilter.start(TransportFilter.java:57)
at org.apache.activemq.transport.WireFormatNegotiator.start(WireFormatNegotiator.java:72)
at org.apache.activemq.transport.TransportFilter.start(TransportFilter.java:57)
at org.apache.activemq.transport.TransportFilter.start(TransportFilter.java:57)
at org.apache.activemq.ActiveMQConnectionFactory.createActiveMQConnection(ActiveMQConnectionFactory.java:258)
... 3 more

This is the very first exception I got when I started to code my first JMS code. Due to wrong Listener Connection (IN red and green (my mistake of taking 9))


package oc.messageQ.olivia;

import java.io.File;
import java.io.FileInputStream;
import java.util.Hashtable;

import javax.jms.Connection;
import javax.jms.ConnectionFactory;
import javax.jms.Destination;
import javax.jms.JMSException;
import javax.jms.MessageProducer;
import javax.jms.Queue;
import javax.jms.QueueConnection;
import javax.jms.QueueConnectionFactory;
import javax.jms.QueueReceiver;
import javax.jms.QueueSender;
import javax.jms.QueueSession;
import javax.jms.Session;
import javax.jms.TextMessage;
import javax.naming.Context;
import javax.naming.InitialContext;

public class IROPSQueueApacheMQProducer {

/**
* @author O Cassandrae
* @param args
*/
public static void main(String[] args) {
MessageProducer producer = null;
Connection conn = null;
Session jmsSession = null;

try {
Hashtable<String, String> env = new Hashtable<String, String>();
env.put(Context.INITIAL_CONTEXT_FACTORY,"org.apache.activemq.jndi.ActiveMQInitialContextFactory");
                        //env.put(Context.PROVIDER_URL,"tcp://PDCHCL-TRAX9:61616");
env.put(Context.PROVIDER_URL,"tcp://PDCHCL-TRAX6:61616");
InitialContext context = new InitialContext(env);
ConnectionFactory cf = (ConnectionFactory) context
.lookup("ConnectionFactory");
System.out.println("ConnectionFactory is ::: "
+ cf.getClass().getName());

if (cf instanceof QueueConnectionFactory) {
conn = ((QueueConnectionFactory) cf).createQueueConnection();
jmsSession = ((QueueConnection) conn).createQueueSession(false,
Session.AUTO_ACKNOWLEDGE);
}
Destination dest = (Destination) context
.lookup("dynamicQueues/GROUNDOPS.PSCS.IROPS.BRIDGE.QUEUE");
System.out.println("dest is ::: " + dest);
            producer = ((QueueSession)jmsSession).createSender((Queue)dest);
            TextMessage message = jmsSession.createTextMessage();
            String AsmSource = "src/resources/IROPSFlightCancelled.xml";
            FileInputStream inputStream;
            byte[] readBytes;
            // build an asm  request
            inputStream = new FileInputStream(AsmSource);
            readBytes = new byte[Integer.valueOf(""+(new File(AsmSource)).length())];
            inputStream.read(readBytes);
            message.setText(new String(readBytes));
            //System.out.println("request: "+new String(readBytes));
            ((QueueSender)producer).send(message);
            //System.out.println("Message sent successfully!!!");
} catch (Exception e) {
e.printStackTrace();
} finally {
try {
if(conn != null){
producer.close();
jmsSession.close();
conn.close();
}
} catch (JMSException e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
}
}
}

Please do start the bat file: C:\Software\apache-activemq-5.5.1\bin\activemq.bat (activemq.bat) before executing the program.

TransientObjectException - Hibernate 2

Transient Object Exception occurs due to two reasons:

1.  The below article will explain it. 

               Transient State

2. It occurs due to passing null as primary key. Check the class you are passing into the table is not null due to any logic. Therefore check the logic of the class it is throwing the exception.

The conversion of the transient state to persistent state in the method getHibernateTemplate();

TransientObjectException - hibernate

org.hibernate.TransientObjectException: object references an unsaved transient instance – save the transient instance before flushing: com.swacorp.pcs.common.model.ServiceLevel
<class name=”com.xxx.A” table=”A” schema=”TESTSCHEMA”>
<id name=”aId” type=”java.lang.Long”>
<column name=”A_ID” precision=”29? scale=”0? />
</id>
…………………..
some more mapping elements
…………………..
…………………..
<many-to-one name=”bId” class=”com.xxx.B” fetch=”select”>
<column name=”B_ID” precision=”29? scale=”0? />
</many-to-one>
…………………..
A is referring to B using a primary key column of bId of B.
In that post I have mentioned that if B is a transient object and you don’t want to persist the value of B to A then just tell the hibernate to ignore that value by saying
update=”false” insert=”false” in the many to one mapping.
But what if you want to persist the value of foreign key in A.
Then the approach is different. You have to make sure that instance B is persistent not transient.
That is if your code says something like
A a = new A();
B b = new B();
a.setB(b);
…..
…..
session.save(a);
you are in trouble. Because B is in transient state. You have to attach b to the session.
There may be other ways of attaching this transient object to session. The approach I am following is simple. I am reading the value of B from the database using Hibernate. That way, hibernate attaches B to session and it is then a persistent object.
That is I do something like
A a = new A();
B b = session.get(B.class, new Long(1));
a.setB(b);
…..
…..
session.save(a);
Hibernate defines and supports the following object states:

  * Transient :- an object is transient if it has just been instantiated using the new operator, and it is not associated with a Hibernate Session. It has no persistent representation in the database and no identifier value has been assigned.

  * Persistent :- a persistent instance has a representation in the database and an identifier value. It might just have been saved or loaded, however, it is by definition in the scope of a Session.

  * Detached :- a detached instance is an object that has been persistent, but its Session has been closed. The reference to the object is still valid, of course, and the detached instance might even be modified in this state. A detached instance can be reattached to a new Session at a later point in time, making it persistent again.

My changes:
HibernateTemplate template = getHibernateTemplate();
   Session session = template.getSessionFactory().getCurrentSession();
   itmAndPaxItmEvntOccrs.paxItm.setSvcLvl((ServiceLevel)session.get(itmAndPaxItmEvntOccrs.paxItm.getSvcLvl().getClass(), new Long(1)));
   template.saveOrUpdate(itmAndPaxItmEvntOccrs.paxItm);

Dont forget to validate your assumptions

There is just so much you can learn from customers (or potential customers) that you should spend at least one day a week talking to them. You'll be amazed at what you can learn.

Technical founders especially love dreaming up cool stuff and getting to coding.

This is a huge mistake.

I recently spoke to one of our customers and asked them what their favourite feature was. I was stunned when they mentioned that sending email when a comment was added was a killer feature for them.

The shock was because we always thought this feature as a commodity feature. I mean, every tool out there has the ability to send email when a comment is added. We had a whole lot of really killer features, but this was one of the key features??

So I probed further.

Turns out that other tools only send out emails for discussions that you are involved in. Our tool sends out emails for everything. Think of a mailing list - you get every email whether you participated or not. Whereas a forum only sends you notification for new replies in the threads where you participated. It was something like that.

We always thought sending email for everything was a limitation. It was on our roadmap to refine it and make it send email selectively.

Well, guess what? Rather than being a limitation, it was actually a feature! And a killer feature for them.

So I called another customer - and this was an important differentiating feature for them too!

In a startup, we make a ton of assumptions. Don't forget to get them validated. Usually what the customer thinks is important is not what you assumed it would be.

Twelve Reasons Why Businesses Fail

From Naeem Zafar's book Seven Steps to a Successful Startup:

1. Solving a problem that most users are not willing to pay to solve your way
2. Thinking that you can do it all by yourself
3. Lacking trust among team members
4. Being overconfident and not questioning yourself
5. Lacking a crisp, singular focus&emdash;Trying to be everything to everyone
6. Marketing myopia
7. Confusing a hobby with a business
8. Pricing incorrectly and not knowing your real competition
9. Failing to properly define your market and customers
10. Not having enough financial resources available
11. Focusing on a market segment too small to sustain you
12. Starting a business for the wrong personal reasons

The difference between a conference and an unconference

One of the things I really like about unconferences is that there is very little "gyan" and a lot of experience sharing. This came home to me when participating in a session on hiring.

The question was: "What strategies do you follow for hiring and building a team?"

This same topic was discussed in a panel in the NASSCOM Product Conclave last year, with some VCs on the panel. In typical panel style, a lot of gyan was thrown around: "Hire only the best people", "Top people create top teams" etc.

That's stating the obvious. I mean which startup doesn't want to hire the best people? Do these VCs believe startups are intentionally hiring poor people because they believe its a better strategy??

The fact is that there are a lot of on the ground challenges to hiring that these panelists are hand waving away with their gyan:

1. Startups are always short of cash and cannot match the compensation in bigger companies
2. A lot of startups hire good guys only for them to leave in 6-12 months
3. Really good guys are hard to find
4. Most employees are not interested in stock options (a lot of Silicon Valley returned VCs talk about stock options... one wonders if they have actually recruited in India)

At the end of the panel, the audience is usually unsatisfied. The panelists tell them what they already know without a word on tackling ground level challenges that they face.

Now compare that with an unconference. Once more this topic came up at the TiE unconference.

Look at the answers this time:

1. Many top people are around in smaller cities, who are not taken by big companies because of a lack of english skills. They are smart and make good startup employees
2. Find a few people who believe in your vision and then complement them with freshers
3. Look at 6 month internships - lots of smart people available as interns
4. Look at your requirements - not all types of work require the best people. Some types of work are repetetive and may just make good employees bored.
5. Look at the attitude and teach the skills

These are strategies that are coming from the experience of people who actually have to tackle these on-the-ground problems. At the end, the participants feel charged up with ideas to take back and implement the next day.

This is really what makes an unconference different for me. The focus on real solutions to real problems is invaluable.

By Siddhu (S3)

3 Oct OCC Meetup Roundup

We had the October 2010 meetup of the Chennai OCC. Tony Aug, VP IT of Sanmina had come down to attend the meet. Once of the big problems startups often have is understanding how large enterprises make decisions. This is really important in the context of B2B enterprise sales. So it was really good to have Tony come down and talk about the executive perspective when purchasing software.

Some points from the meetup -

• The #1 reason for purchase in a recessionary environment right now is to cut costs. If you can show substantial cost reductions, then that is a big win.
• Many cloud/SaaS startups dont understand the security and privacy needs of enterprises. Eg: A company like Sanmina with operations in 20 countries need to run only on certified infrastructure that comply to EU security and privacy regulations, HIPAA, and a host of other regulations. Startups right now cannot guarantee all this
• Pricing: $10/user a month sounds reasonable, but multiply it by 50,000 users and it becomes huge. So you will need to find a way to get it into the team on one pricing model and then switch to another pricing model (possibly a flat price) when rolling out to the whole organization
• Dont forget about switching costs. Once an organization has spent tons of money on training and integration, it is unlikely that they will switch over, even to a superior product, unless there is a really easy path for them to do so
• Subscription pricing is great because it allows organizations to get started without expensive capital expenditure. The old model was to buy expensive servers, install expensive software, training, integration and then you could look to get RoI. The new model is to subscribe and cancel if it doesn't work out
• You can expect a small team to pay by credit card, but larger orders will need a purchase order and go via the purchasing department. Once that happens a lot of other stakeholders will look at it and question the need for the purchase. So you will have to answer to these other stakeholders too.
• Lock-in is important in a subscription service. Most companies will want to be sure that they can get their data out at any time
• Sanmina is a huge believer in open source - cuts costs and there is a community to help. For mission critical systems they pay for commercial support
• Escrow: Larger companies will usually ask startups that the source code be put in escrow. In case the startup closes down, the company gets access to the source.
• Using tools like Linkedin, it is really easy nowadays to find a path to get to a decision maker in any company

By Siddhi (S3)

Sales Stories!

This post is the first in a series of stories on selling.

This stories in this post are from the book Hope is not a strategy.

On targeting the pain point

    Early in my sales career as an account manager, I was in the back of the room while one of my product specialists was presenting a system. We were extremely proud of the functionality and about halfway through the presentation, our product rep put up a slide and said, "Now, if you were a hospital, you'd really like this feature."

    What!? I was utterly stunned. I thought to myself, "They're not a hospital-they're a bank! They're never going to be a hospital; they'll always be a bank!"

    How preposterous, the idea of showing a hospital feature to a bank.

    But this is no less preposterous than showing somebody a solution to a problem they don't have-or a feature for which they have no need. You might as well show a hospital feature to a bank.


On trust

    The evaluation process was grueling. Two vendors were pitted against each other and were required to do detailed benchmark implementations in the evaluation of their products. It was an exhausting, detailed process with perhaps twelve to twenty people on each vendor's side.

    In the end, we won by a couple of points, but we broadened that advantage into a business partnership. Their executives met our chairman. We had a corporate visit. We continued to talk about implementation and support and widened the gap. We actually reached a measure of business partnership before they signed the contract.

    At dinner the night before the contract signing for multiple financial systems, the client said, "Oh by the way, do you have a fixed asset system?"

    "Certainly"

    "How much is it?"

    "Seventy-five thousand dollars."

    "Add it to the proposal."

    I thought, "They just bought a system sight unseen. How could they have required such detailed evaluation in the beginning and now they have bought a system they never even looked at?"

    But they went further.

    They added the graphic user interface for all the systems, which they knew was in the prototype phase. There were no references for this three-hundred-thousand-dollar product at that time. The opportunity went from six hundred thousand dollars to almost a million dollars, and forty percent of it was on products they had never seen.

    How could they evaluate in the beginning in such detail and buy products sight-unseen now?

    The answer is trust. They trusted our products based on the ones they had evaluated. If those worked, then these must. They trusted us as a company because they had met our top executives, they'd been to our headquarters, and they knew we were a stable industry leader with a good track record of delivery. And they trusted us personally because we understood their business and we had become friends.

By Siddhi (S3)

Searching for design Tool?

You are trying to choose a tool for creating UML diagrams of all flavours? And usability is a major criteria for you, but you'd still take more power with a steeper learning curve and be happy? Free (as in beer) would be nice? Still you'd be willing to pay if the tool's worth it. What should I be using?

Some context: Recently for graduate school I researched UML tools for usability and UML comprehension in general for an independent project. I also model/architect for a living.

The previous posts have too many answers and not enough questions. A common misunderstanding is that UML is about creating diagrams. Sure, diagrams are important, but really you are creating a model. Here are the questions that should be answered as each vendor product/solution does some things better than others. Note: The listed answers are my view as the best even if other products support a given feature or need.

• Are you modeling or drawing? (Drawing - ArgoUML, free implementations, and Visio)
• Will you modeling in the future? (For basic modeling - Community editions of pay products)
• Do you want to formalize your modeling through profiles or meta-models? OCL? (Sparx, RSM, Visual Paradigm)
• Are you concerned about model portability, XMI support? (Sparx, Visual Paradigm, Altova)
• Do you have an existing set of documents that you need to work with? (Depends on the documents)
• Would you want to generate code stubs or full functioning code?(Visual Paradigm, Sparx, Altova)
• Do you need more mature processes such as use case management, pattern creation, asset creation, RUP integration, etc? (RSA/RSM/IBM Rational Products) 

Detailed Examples: IBM Rational Software Architect did not implement UML 2.0 all the way when it comes to realizes type relationships when creating a UML profile, but Visual Paradigm and Sparx got it right.
Ok, that was way to detailed so a simpler example would be ArgoUML has no code generation features and focuses on drawing more than the modeling aspect of UML.
Sparx and Visual Paradigm do UML really well and generate code well, however, hooking into project lifecycles and other process is where RSM/RSA is strong.
Watch out for closed or product specific code generation processes or frameworks as you could end up stuck with that product.

This is a straight brain dump so a couple details may not be perfect, however, this should provide a general map to the questions and solutions to looking into.

NEW - Found a good list of many UML tools with descriptions. Wiki UML Tool List

Source