6.1.4 Deleting process instance execution logs

(a) Deleting execution logs individually

The procedure to delete a process instance execution log one by one using screen is as follows:

1. Search the process instances.

2. From the search results list, select the process instance execution log you want to delete.

3. Click the [Delete] button.

   A dialog box confirming deletion of the process instance execution log is displayed.

   If the deletion target includes a process instance in Executing status, a dialog box confirming deletion of the execution log for the process instance that is executing is displayed.

4. Click the [OK] button.

   The process instance execution log is deleted, and a dialog box showing the result of deleting the process instance execution log is displayed.

5. Click the [OK] button.

   The list of search results is updated.

(b) Deleting in batch mode

The procedure to delete process instance execution log in batch mode using screen is as follows:

1. Search the process instances.

2. From the search results list, select the process instance execution log you want to delete.

   Use the [Ctrl] or [Shift] key to select multiple process instances.

3. Click the [Delete] button.

   A dialog box confirming deletion of the process instance execution log is displayed.

   If the deletion target includes a process instance in Executing status, a dialog box confirming deletion of the execution log for the process instance that is executing is displayed.

4. Click the [OK] button.

   The process instance execution log is deleted, and a dialog box showing the result of deleting the process instance execution log is displayed.

5. Click the [OK] button.

   The list of search results is updated.

   Important note

   If batch deletion fails, none of the selected process instances is deleted.

(a) Deleting execution logs individually

The procedure to delete process instance execution log individually using command is as follows:

1. Determine the HCSC server to be accessed.

2. Use the cscpiselect command to search the process instance to be deleted.

3. From the search results, identify the process ID of the process instance whose execution log you wish to delete.

4. Execute the following command to delete the execution log:

   • When directly specifying the where clause

     cscpidelete -user <Login user ID> -pass <Login password> -csc <HCSC server name> -where <where clause>#

   Note #

   Specify the deletion condition for the process instance whose execution log you are deleting.

   • When using a file in which the where clause is specified

     cscpidelete -user <Login user ID> -pass <Login password> -csc <HCSC server name> -wherefile <file name in which the where clause is specified>#

   Note #

   Specify the deletion condition for the process instance whose execution log you are deleting.

   • When specifying a file containing a list of IDs

     cscpidelete -user <Login user ID> -pass <Login password> -csc <HCSC server name> -idlist <file name in which a list of ID is specified>

(Example) Enter command using following condition to delete process instance execution log.

• Process ID to be deleted: J2EEServer_010209010042_XXXPurchaseRegisterService_1142479241203_760331597_2115

• HCSC server to be accessed: HCSC

Command

• When the -wherefile option is specified

  cscpidelete -user <Login user ID> -pass <Login password> -csc HCSC -wherefile delete_where.txt

  Contents of delete_where.txt

  ProcessID = 'J2EEServer_010209010042_XXXPurchaseRegisterService_1142479241203_760331597_2115'

• When -idlist option is specified

  cscpidelete -user <Login user ID> -pass <Login password> -csc HCSC -idlist delete_id_list.txt

  Contents of delete_id_list.txt

  J2EEServer_010209010042_XXXPurchaseRegisterService_1142479241203_760331597_2115

(b) To delete execution logs in a batch by specifying the business process definition name of the process instances

The procedure to specify business process definition name of process instance using command and deleting execution log in batch mode is as follows:

1. Determine the HCSC server to be accessed.

2. Execute the following command to delete the execution log:

   • When directly specifying the where clause

     cscpidelete -user <Login user ID> -pass <Login password> -csc <HCSC server name> -where <where clause>#

   Note #

   Specify the SQL condition for the column name ProcessDefinitionName.

   • When using a file in which the where clause is specified

     cscpidelete -user <Login user ID> -pass <Login password> -csc <HCSC server name> -wherefile <file name in which the where clause is specified>#

   Note #

   Specify the SQL condition for the column name ProcessDefinitionName.

   • When directly specifying a business process definition name

     cscpidelete -user <Login user ID> -pass <Login password> -csc <HCSC server name> -processname <Business process definition name>

(Example) Enter command using following condition to delete process instance execution log in batch mode.

• Business process definition name: Hello business process

• HCSC server to be accessed: HCSC

Command

• When the -wherefile option is specified:

  cscpidelete -user <Login user ID> -pass <Login password> -csc HCSC -wherefile delete_where.txt

  Contents of delete_where.txt

  ProcessDefinitionName=' Hello business process '

• When the -processname option is specified:

  cscpidelete -user <Login user ID> -pass <Login password> -csc HCSC -processname Hello business process

(c) To delete execution logs in a batch by specifying start date and time of process instances

The procedure to specify start date and time of process instance using command and deleting execution log in batch mode is as follows:

Important note

The start date and time to be specified for deleting execution logs differs as per the hour system used in the option specified in the command.

• When specified in the file in which the -where option or the where clause of the -wherefile option is specified: Greenwich Mean Time (GMT)

• When specified in the -date option: Default time zone

1. Determine the HCSC server to be accessed.

2. Execute the following command to delete the execution log:

   • When directly specifying the where clause

     cscpidelete -user <Login user ID> -pass <Login password> -csc <HCSC server name> -where <where clause>#

   Note #

   Specify the SQL condition for the column name StartTime.

   • When using a file in which the where clause is specified

     cscpidelete -user <Login user ID> -pass <Login password> -csc <HCSC server name> -wherefile <file name in which the where clause is specified>#

   Note #

   Specify the SQL condition for the column name StartTime.

   • When directly specifying a start time range

     cscpidelete -user <Login user ID> -pass <Login password> -csc <HCSC server name> -date <inquiry start date and time>#,<inquiry end date and time>

   Note #

   Inquiry start date/time can be omitted.

(Example) Enter the command using the following conditions to delete process instance execution logs in batch mode:

• Start date and time range: Before 2008/10/09 14:59:59 (Greenwich Mean Time)

• HCSC server to be accessed: HCSC

Command

• When the -wherefile option is specified:

  cscpidelete -user <Login user ID> -pass <Login password> -csc HCSC -wherefile delete_where.txt

  For contents of delete_where.txt

  StartTime <= '2008-10-09 14:59:59'

• When the -date option is specified

  cscpidelete -user <Login user ID> -pass <Login password> -csc HCSC -date ,20081009235959#

Note #

By default, Japan time is used for time zone in -date option.

(Example) Enter the command using the following conditions to delete process instance execution logs in batch mode:

• Start date and time range: 2008/10/08 15:00:00 - 2008/10/09 14:59:59(Greenwich Mean Time)

• HCSC server to be accessed: HCSC

Command

• When the -wherefile option is specified

  cscpidelete -user <Login user ID> -pass <Login password> -csc HCSC -wherefile delete_where.txt

  Contents of delete_where.txt

  StartTime BETWEEN '2008-10-08 15:00:00' AND '2008-10-09 14:59:59'

• When the -date option is specified

  cscpidelete -user <Login user ID> -pass <Login password> -csc HCSC -date 20081009000000#,20081009235959#

Note #

By default, Japan time is used for time zone in -date option.

(d) To delete execution logs in a batch by specifying the status of the process instances

The procedure to use commands to specify the status of the process instances and delete execution logs in batch mode is as follows:

1. Determine the HCSC server to be accessed.

2. Execute the following command to delete the execution log:

   • When directly specifying the where clause

     cscpidelete -user <Login user ID> -pass <Login password> -csc <HCSCserver name> -where <where clause>#

   Note #

   Specify the SQL condition for the column name State.

   • When using a file in which the where clause is specified

     cscpidelete -user <Login user ID> -pass <Login password> -csc <HCSC server name> -wherefile <file name in which the where clause is specified>#

   Note #

   Specify the SQL condition for the column name State.

   • When directly specify the status

     cscpidelete -user <Login user ID> -pass <Login password> -csc <HCSC server name> -status {executing | completed | faulted | error | all}#

   Note #

   Specify one of these items.

(Example) Enter the command using the following conditions to delete process instance execution logs in batch mode:

• Status of process instance: Executing

• HCSC server to be accessed: HCSC

Command

• When the -wherefile option is specified

  cscpidelete -user <Login user ID> -pass <Login password> -csc HCSC -wherefile delete_where.txt

  Contents of delete_where.txt

  State='Executing'

Note

Note that, if you specify a condition other than State='Completed', be careful as process instances except for completed process instances get deleted.

• When the -status option is specified

  cscpidelete -user <Login user ID> -pass <Login password> -csc HCSC -status executing

(e) To delete execution logs in a batch by combining the business process definition name, start date and time, and status of the process instances

The procedure to specify the after combining business process definition name of process instances, start date and time, and status using command, and deleting execution logs in a batch mode is as follows:

Important note

The start date and time to be specified for deleting execution logs differs as per the hour system used in the option specified in the command.

• When specified in the file in which the -where option or the where clause of the -wherefile option is specified: Greenwich Mean Time (GMT)

• When specified in the -date option: Default time zone

1. Determine the HCSC server to be accessed.

2. Execute the following command to delete the execution log:

   • When directly specifying the where clause

     cscpidelete -user <Login user ID> -pass <Login password> -csc <HCSC server name> -where <where clause>#

   Note #

   Specify the SQL condition for the column names ProcessDefinitionName, StartTime, and State.

   • When using a file in which the where clause is specified

     cscpidelete -user <Login user ID> -pass <Login password> -csc <HCSC server name> -wherefile <file name in which the where clause is specified>#

   Note #

   Specify the SQL condition for the column names ProcessDefinitionName, StartTime, and State.

   • When directly specifying the business process definition name, start time range and status:

     cscpidelete -user <Login user ID> -pass <Login password> -csc <HCSC server name> -processname <Business process definition name> -date <inquiry start date and time>#1,<inquiry end date and time> -status {executing | completed | faulted | error | all}#2

   Note #1

   Inquiry start date/time can be omitted.

   Note #2

   Specify one of these items.

(Example) Enter the command using the following conditions to delete process instance execution logs in batch mode:

• Business process definition name: Hello business process

• Start date and time range: before 2008/10/09 14:59:59(Greenwich Mean Time)

• Status of process instance: Executing

• HCSC server to be accessed: HCSC

Command

• When the -wherefile option is specified

  cscpidelete -user <Login user ID> -pass <Login password> -csc HCSC -wherefile delete_where.txt

  Contents of delete_where.txt

  ProcessDefinitionName='Hello business process' AND StartTime <= '2008-10-09 14:59:59' AND State='Executing'

• When the -processname option, -date option, and -status option are specified

  cscpidelete -user <Login user ID> -pass <Login password> -csc HCSC -processname Hello business process -date ,20081009235959# -status executing

  Note #

  By default, Japan time is used for time zone in -date option.

(f) To delete the execution logs for all process instances in a batch

The procedure to use commands to delete all process instance execution logs in batch mode is as follows:

1. Determine the HCSC server to be accessed.

2. Execute the following command to delete the execution log:

   cscpidelete -user <Login user ID> -pass <Login password> -csc <HCSC server name> -force

(Example) Enter the command using the following conditions to delete all process instance execution logs in batch mode:

• HCSC server to be accessed: HCSC

Command

cscpidelete -user <Login user ID> -pass <Login password> -csc HCSC -force

(g) To delete execution logs in a batch by specifying the deletion process time

If the quantity of process instance execution logs that you want to delete is large, long period will be required for completing all the deletion processes. You can limit time required for batch deletion process so that deletion process will not affect request process.

You can specify the elapsed time till the deletion process is stopped in seconds in-canceltimer option of cscpidelete command. With this you can stop the deletion process if the time after commit is exceeds the specified time and the deletion process after this point will not be executed.

(Example)

In the following example, data saved till 31/3/2005 is deleted in 600 seconds

cscpidelete -csc CSCSrv1 -date ,20050331235959 -canceltimer 600

For details on cscpidelete command, see "cscpidelete (Deleting the process instance execution log)" in "Service Platform Reference Guide".

(a) Steps to save execution log

Prerequisites when storing data are as follows:

• Confirm that reception of all the HCSC servers is stopped.

• Confirm that there is no request is in progress.

• Do not perform command of Service Coordinator.

• Execute storage process using memory (Java program).

Storage process of data is as follows:

1. Connect to database (HiRDB ) using JDBC driver.

   Be careful about following points when connecting

   • It is recommended to set LONGVARBINARY_ACCESS of user property when connecting as LOCATOR.

     When you set REAL in LONGVARBINARY_ACCESS, set maximum value appropriate for HiRDB_for_Java_MAXBINARYSIZE.

     When HiRDB_for_Java_MAXBINARYSIZE is not set, since 147,483,639 byte memory is allocated in data acquisition of Variable value of CSCBP_<Cluster name>_MSG_VARIABLE table, there is a possibility that java.lang.OutOfMemoryError may occur.

     Also, when setting value of HiRDB_for_Java_MAXBINARYSIZE is less than maximum length of data stored in VariableValue of CSCBP_<Cluster name>_MSG_VARIABLE table, exceeded data is omitted and there is possibility that system may not operate normally.

   • Disabling automatic commit is recommended.

     If it is not disabled, in case of an error in storage process, the system may not operate normally.

2. Identify the process instances to be stored.

   For storing the process instances which are stored in RD area to be deleted or when status of process instance is "Executing" or "Error", issue the following SQL and acquire a list of Process instance identifier to be stored.

   SELECT ProcessID FROM CSCBP_<Cluster name>_PROCESS
           IN ('<RD area name 1 which you want to delete>,<RD area name 2 which you want to delete>,...')
           WHERE State IN ('Executing', 'Error')
          WITHOUT LOCK NOWAIT;

3. Change the storage location of the process instances to be stored.

   Consider Process instance identifier acquired in procedure 2.as key and change storage location of the entire following table.

   • CSCBP_<Cluster name>_PROCESS

   • CSCBP_<Cluster name>_ACTIVITY

   • CSCBP_<Cluster name>_LINK

   • CSCBP_<Cluster name>_CORRELATIONSET

   • CSCBP_<Cluster name>_STR_VARIABLE

   • CSCBP_<Cluster name>_NUM_VARIABLE

   • CSCBP_<Cluster name>_BOOL_VARIABLE

   • CSCBP_<Cluster name>_MSG_VARIABLE

   • CSCBP_<Cluster name>_MSG_RELATION

   The procedure to change storage location is as follows. When you have acquired multiple Process instance identifiers, perform process (i)-(iii) multiple times.

   (i)Acquire entire data of target Process instance identifier

   Issue following SQL and acquire entire data to be stored.

   SELECT * FROM CSCBP_<Cluster name>_PROCESS
       IN ('<RD area name 1 which you want to delete>,<RD area name 2 which you want to delete>,...')
   WHERE ProcessID = <Process instance identifier> WITHOUT LOCK NOWAIT;
    
   SELECT * FROM CSCBP_<Cluster name>_ACTIVITY
       IN ('<RD area name 1 which you want to delete>,<RD area name 2 which you want to delete>,...')
   WHERE ProcessID = <Process instance identifier> WITHOUT LOCK NOWAIT;
    
   SELECT * FROM CSCBP_<Cluster name>_LINK
       IN ('<RD area name 1 which you want to delete>,<RD area name 2 which you want to delete>,...')
   WHERE ProcessID = <Process instance identifier> WITHOUT LOCK NOWAIT;
    
   SELECT * FROM CSCBP_<Cluster name>_CORRELATIONSET
       IN ('<RD area name 1 which you want to delete>,<RD area name 2 which you want to delete>,...')
   WHERE ProcessID = <Process instance identifier> WITHOUT LOCK NOWAIT;
    
   SELECT * FROM CSCBP_<Cluster name>_STR_VARIABLE
       IN ('<RD area name 1 which you want to delete>,<RD area name 2 which you want to delete>,...')
   WHERE ProcessID = <Process instance identifier> WITHOUT LOCK NOWAIT;
    
   SELECT * FROM CSCBP_<Cluster name>_NUM_VARIABLE
       IN ('<RD area name 1 which you want to delete>,<RD area name 2 which you want to delete>,...')
   WHERE ProcessID = <Process instance identifier> WITHOUT LOCK NOWAIT;
    
   SELECT * FROM CSCBP_<Cluster name>_BOOL_VARIABLE
       IN ('<RD area name 1 which you want to delete>,<RD area name 2 which you want to delete>,...')
   WHERE ProcessID = <Process instance identifier> WITHOUT LOCK NOWAIT;
    
   SELECT * FROM CSCBP_<Cluster name>_MSG_VARIABLE
       IN ('<RD area name 1 which you want to delete>,<RD area name 2 which you want to delete>,...')
   WHERE ProcessID = <Process instance identifier> WITHOUT LOCK NOWAIT;
    
   SELECT * FROM CSCBP_<Cluster name>_MSG_RELATION
       IN ('<RD area name 1 which you want to delete>,<RD area name 2 which you want to delete>,...')
   WHERE ProcessID = <Process instance identifier> WITHOUT LOCK NOWAIT;

   (ii)Update value of SplitKey column and re-enter in the table.

   Change value of SplitKey column in such a way that next it is allocated to next RD area to be deleted (For example: when breakup of RD area is per 1 month and current target deletion is 12 months ('12'), change it to 1 ('01'), issue INSERT statement.

   For this, issue following SQL for all acquired data, you need to re-register in such a way that data to be stored is allocated to other RD area. Issue INSERT statement only with number records acquired in SELECT statement of procedure (i)

   INSERT INTO CSCBP_<Cluster name>_PROCESS
   (ProcessID,ProcessDefinitionName,ProcessDefinitionVersion,State,ChangeCount, 
   StartTime,EndTime,CSCServerName,SplitKey)
   VALUES (<Data except for data in SplitKey example acquired in (i)>, '01');
    
   INSERT INTO CSCBP_<Cluster name>_ACTIVITY
   (ProcessID,ActivityNumber,ActivityDefinitionName,ActivityKind,ParentNumber,
   ScopeNumber,RepeatNumber,State,FirstChildNumber,LastChildNumber,StartTime,EndTime,
   WaitTime,SplitKey)
   VALUES (<Data except for data in SplitKey example acquired in (i)>, '01');
    
   INSERT INTO CSCBP_<Cluster name>_LINK
   (ProcessID,LinkDefinitionName,RepeatNumber,State,SplitKey)
   VALUES (<Data except for data in SplitKey example acquired in (i)>, '01');
    
   INSERT INTO CSCBP_<Cluster name>_CORRELATIONSET
   (ProcessID,ProcessDefinitionName,ProcessDefinitionVersion,ScopeDefinitionName,
   ScopeNumber,CorrelationSetName,CorrelationSetValue,State,SplitKey)
   VALUES (<Data except for data in SplitKey example acquired in (i)>, '01');
    
   INSERT INTO CSCBP_<Cluster name>_STR_VARIABLE
   (ProcessID,ScopeNumber,VariableName,VariableValue,SplitKey)
   VALUES (<Data except for data in SplitKey example acquired in (i)>, '01');
    
   INSERT INTO CSCBP_<Cluster name>_NUM_VARIABLE
   (ProcessID,ScopeNumber,VariableName,VariableValue,SplitKey)
   VALUES (<Data except for data in SplitKey example acquired in (i)>, '01');
    
   INSERT INTO CSCBP_<Cluster name>_BOOL_VARIABLE
   (ProcessID,ScopeNumber,VariableName,VariableValue,SplitKey)
   VALUES (<Data except for data in SplitKey example acquired in (i)>, '01');
    
   INSERT INTO CSCBP_<Cluster name>_MSG_VARIABLE
   (ProcessID,ScopeNumber,VariableName,VariableValue,SplitKey)
   VALUES (<Data except for data in SplitKey example acquired in (i)>, '01');
    
   INSERT INTO CSCBP_<Cluster name>_MSG_RELATION
   (MessageID,ProcessID,ActivityNumber,SendOrReceive,
   AsynchronousMessagingID,SplitKey)
   VALUES (<Data except for data in SplitKey example acquired in (i)>, '01');

   Note

   • When you are updating 1-9 months, put 0 before the number as shown in "01".

   • Do not change value other than value in SplitKey column.

   • Match the value of SplitKey column of similar Process instance identifier.

     Note that, when you invoke subsequent process after re-executing stored process instance and specifying correlation set, the value of SplitKey column in data newly stored in database is the value similar to value of SplitKey which is INSERT in this procedure and not the value in "month (GMT) when process instance was started"

   (iii) Commit transaction

   If you execute large number of processes when automatic commit is disabled, there is possibility that shortage of resources may occur. Issue commit of transaction periodically (Example: per 100 Process instance identifier), release resources.

4. Abort connection.

(b) Procedure to delete execution log

The deletion procedure per RD area is as follows:

1. Block RD area wherein the data to be deleted is stored.

   Execute following command for RD area for index and RD area for data both.

   pdhold -r <RD area name>

2. Unloading the data stored in RD area which is to be deleted.

   This process is optional. It is performed only when you want to take back up of the data stored in RD area.

   Perform unload per table as shown in following format.

   pdrorg -k unld -t <Table name>#1 -r <RD area name for data> <Control information file name>#2

   Note #1

   Specify following names in table name.

   CSCBP_<Cluster name>_PROCESS

   CSCBP_<Cluster name>_ACTIVITY

   CSCBP_<Cluster name>_LINK

   CSCBP_<Cluster name>_CORRELATIONSET

   CSCBP_<Cluster name>_STR_VARIABLE

   CSCBP_<Cluster name>_NUM_VARIABLE

   CSCBP_<Cluster name>_BOOL_VARIABLE

   CSCBP_<Cluster name>_MSG_VARIABLE

   CSCBP_<Cluster name>_MSG_RELATION

   Note #2

   Describe following contents in specified control information file.

   In HiRDB/Single Server

   unload <path of file that unloads data of each table>

   In HiRDB/Parallel Server

   unload <Backend server name>:<path of file that unloads data of each table>

   For details on storing the data unloaded in pdrorg to table and method of fixing, see description of pdrorg output binary format file in "HiRDB Version 10 Command reference (For Windows(R))" or "HiRDB Version 10 Command reference (For UNIX(R)".

3. By deleting the data stored in RD area, load data with 0 cases. Loading method is as follows. The index information matching with data contents is also loaded. See contents of error information file which is output and confirm that error has not occurred.

   pdload -d CSCBP_<Cluster name>_PROCESS <Control information file name>
   pdload -d CSCBP_<Cluster name>_ACTIVITY <Control information file name>
   pdload -d CSCBP_<Cluster name>_LINK <Control information file name>
   pdload -d CSCBP_<Cluster name>_CORRELATIONSET <Control information file name>
   pdload -d CSCBP_<Cluster name>_STR_VARIABLE <Control information file name>
   pdload -d CSCBP_<Cluster name>_NUM_VARIABLE <Control information file name>
   pdload -d CSCBP_<Cluster name>_BOOL_VARIABLE <Control information file name>
   pdload -d CSCBP_<Cluster name>_MSG_VARIABLE <Control information file name>
   pdload -d CSCBP_<Cluster name>_MSG_RELATION <Control information file name>

   In specified control information file, describe following contents. Arrange for the file with data in which there are 0 cases (Contents are empty), to be specified in <Data file path>.

   For HiRDB/Single Server

   source <RD area name for data> <Data file path>

   For HiRDB/Parallel Server

   source <RD area name for data> <Backend server name>:<Data file path>

4. Release the lock in RD area

   Execute following command and release the lock of RD area which was implemented in step 1.

   pdrels -r <RD area name>

Issue following SQL and confirm if RD area units are deleted correctly with this procedure:

SELECT count(*) FROM CSCBP_<Cluster name>_PROCESS IN ('<RD area name 1 which you want to delete>,<RD area name 2 which you want to delete>,...') WITHOUT LOCK NOWAIT;