Chat now with support
Chat with Support

SharePlex 10.2 - Administration Guide

About this Guide Conventions used in this guide Revision History Overview of SharePlex Run SharePlex Run multiple instances of SharePlex Execute commands in sp_ctrl Set SharePlex parameters Configure data replication Configure replication to and from a container database Configure named queues Configure partitioned replication Configure replication to a change history target Configure a replication strategy Configure DDL replication Configure error handling Configure data transformation Configure security features Start replication on your production systems Monitor SharePlex Prevent and solve replication problems Repair out-of-sync data Tune the Capture process Tune the Post process Recover replication after Oracle failover Make changes to an active replication environment Apply an Oracle application patch or upgrade Back up Oracle data on the source or target Troubleshooting Tips Appendix A: Peer-To-Peer Diagram Appendix B: SharePlex environment variables

Solve compare command errors

If you are having trouble with any of the compare or repair commands, refer to the following for assistance.

If the issue you are experiencing is not listed in this documentation, search the SharePlex Knowledge Base at:

https://support.quest.com.

The Knowledge Base provides filtering options and links to other resources that can help you use and troubleshoot SharePlex.

Problem Description Solution
Oracle error 904

Error 904 calling oexec in de_select_prepare_to_fetch.

A comparison failed because the source and target tables being compared are structurally different. The compare and repair commands do not detect and repair out-of-sync conditions caused by unsynchronized DDL operations or tables that are not structurally identical.

Do a DESCRIBE on both tables. It probably will show that the tables do not have an equal number of columns or the data types are different. After you correct the DDL problem, you can run a repair to resynchronize the values in the rows.

"Too many users” error

Can not add DataEquator queue reader que_TOOMANYUSERS: User table is full.

The maximum number of processes reading from and writing to the SharePlex queues was exceeded. No more than 20 processes can read from and write to the post queue at the same time, including the replication processes and the compare and repair processes. An error is most likely to occur when a repair option is used, because a repair accesses the queue much longer than a comparison without a repair.

There is no workaround or way to adjust the limit, nor is there a way to determine how many compare processes can run concurrently without exceeding the limit.

TIP: To compare multiple tables at the same time, without being restricted by process limitations, use the compare using command. To limit the tables being compared, create a new configuration containing only the ones that you want to compare, and use it for the comparison. (Do not activate that configuration.) All of the tables are compared with one compare using process.

A client process fails to die

When a sp_desvr server process dies, the associated sp_declt client processes usually die. If a process does not die, you can kill it.

For more information, see Kill compare processes.
A server process fails to die When you kill a sp_declt client process or it dies on its own — or if the sp_desvr server process has not communicated with the client — the sp_desvr server process usually exits after a certain amount of time, which is controlled by the SP_DEQ_TIMEOUT parameter. For more information, see Kill compare processes.

Kill compare processes

To kill a client process

If you need to kill a sp_declt client process, and there are multiple compares running, you can determine which one to kill in one of the following ways:

  • By viewing the sp_declt log file — In the file, look at the Session IDs of the sp_declt processes and find the one that matches the PID of the sp_desvr process that died. That is the sp_declt process to kill. The sp_declt Session ID is the same as the PID of the associated sp_desvr process.
  • By viewing the Event Log — The Event Log records the startup of each sp_declt client process and its PID. A subsequent entry in the log records the compare log file to which the process is writing. Within the compare log file’s name in that entry is the PID of the server process. For example, in the following sample entry, the sp_declt process PID is 2450. The process writes to compare log ../o734v32a_declt- 1228-01.log. The 1228 is the PID of the server process.

    05/04/01 17:01 Process launched: sp_declt (for o.o734v32a-o.o734v32a- 87056 queue all) [pid = 2450]

    05/04/01 17:01 Notice: sp_declt(deq) (for o.o734v32a-o.o734v32a-87056 queue all) Opened DataEquator session log file /u10/julia30014/var7/ log/o734v32a_declt-1228-01.log

    You can search the log file names for the server process that died, and look for the client process associated with that log file to determine the correct PID to kill.

On Windows systems, the logs also record the startup of the associated sp_cop process.

To kill a server process

If you need to kill a sp_desvr server process when a sp_declt client process dies, look in the Event Log to find out which log the sp_declt client process was writing to. The Event Log records the startup of each client process and its PID. A subsequent entry in the log records the compare log file to which the process is writing. Within the compare log file’s name in that entry is the PID of the server process. For example, in the following sample entry, the sp_declt process PID is 2450. The process writes to log ../o734v32a_declt-1228-01.log. The 1228 is the PID of the server process, and that is the process to kill.

05/04/01 17:01 Process launched: sp_declt (for o.o734v32a-o.o734v32a- 87056 queue all) [pid = 2450]

05/04/01 17:01 Notice: sp_declt(deq) (for o.o734v32a-o.o734v32a-87056 queue all) Opened DataEquator session log file /u10/julia30014/var7/ log/o734v32a_declt-1228-01.log

On Windows systems, the logs also record the startup of the associated sp_cop process.

Solve other problems

This section reviews solutions to other replication problems.

If the issue you are experiencing is not listed in this documentation, search the SharePlex Knowledge Base at:

https://support.quest.com.

The Knowledge Base provides filtering options and links to other resources that can help you use and troubleshoot SharePlex.

SharePlex does not run on a Windows system

SharePlex uses the MKS Toolkit® (also known as NuTCRACKER) operating environment from Parametric Technology Corporation (PTC) to run on Windows systems. If the NuTCRACKER service is stopped or disabled, or if the NuTCRACKER files have been removed or relocated, there will be errors when you try to run SharePlex.

Solution:

  1. Run the Windows Task Manager, then click the Processes tab.
  2. Check to see if the NuTCRACKER service is running. The process name is nutsrvx, where x is the version of the NuTCRACKER software.
  3. If the process is not running, start the NuTCRACKER service by using the Services panel of the Administrative Tools Control Panel.

    • If you do not see the NuTCRACKER service, re-install SharePlex.
    • If the NuTCRACKER service started, but SharePlex still returns errors, check to see if the NuTCRACKER files were relocated. To determine the correct installed location, look at the following in the Windows Registry:

      HKEY_LOCAL_MACHINE\SOFTWARE\Data Focus\Runtime

    • The InstallDir string in the right pane of the Runtime node shows the correct location for the files. Search for the MKS Toolkit folder and restore the files to the location specified in the InstallDir entry.

If you cannot locate the files or cannot restore them to the correct location, do the following:

  1. Stop the SharePlex and NuTCRACKER services, if running.

  2. Run regedit to open the Registry Editor.

  3. Delete the Data Focus and Mortice Kern Systems registry folders under HKEY_LOCAL_MACHINE\SOFTWARE\.
  4. Close the Registry Editor.
  5. Restart the system.
  6. Reinstall SharePlex. Make certain to reinstall SharePlex in the same location, and make certain to install the NuTCRACKER component when prompted.
  7. Restart the system to make the new NuTCRACKER environment effective.

A “Can’t unlink file” error occurs on Windows systems

A nuisance error similar to the ones below sometimes occur on Windows systems. The files eventually unlink.

Text file busy Unlinking file: 'r:\splex2102/rim/o.SERV+C+0.0000000

Or...

System call error: sp_ordr.exe(osp) (for o.SERV queue o.SERV) Text file busy 17003 - Can't unlink file R:\Splex2100/state/o.SERVlog_ sp_ordr.30

Common connection errors

The following are solutions to common errors when starting sp_ctrl, or with forming a connection with the host, port or [on host] commands in sp_ctrl.

Explanation of connection error messages
Error Cause Solution
Host unknown: cannot form connection Appears when either the host command or [on host] option is issued. Verify that the system to which you want to connect is running and that you are using the correct system name.
Network unreachable The network is down. Find out how long the network administrator expects it to last. If the downtime could cause the SharePlex queues to exceed their disk space, take measures to avoid having to resynchronize the data. For more information, see How to resolve disk space shortage.

Export cannot connect to import on hostname: timeout waiting for ack

Export cannot connect to the target because its connection was timed out by the network configuration. This can occur when there is little replication activity and the network has a timeout setting.

Set the SP_XPT_KEEPALIVE parameter to 1. This setting tells the Export process to send a "hello" message to Import at regular intervals to prevent the TCP timeouts.

User is not authorized as a SharePlex user -- check /etc/group You do not have user permissions to execute the operation. SharePlex users must be listed in the /etc/group file (Unix and Linux) or in the Users list (Windows) under one of the SharePlex user groups: SharePlex Admin, spoper, spview.
unauthorized connection attempt from host hostname. net A connection from a remote machine was denied because its name is not listed in the auth_hosts file. See the error message for the name of the system. To allow that system to connect to the sp_cop on the local system, add its name to the auth_hosts file.

Common command errors

Error Cause Solution
Deactivate/flush a nonactive datasource You are attempting to flush a configuration that is not active. None required.
Bad routing specification The syntax in the routing map is incorrect. For more information, see Routing specifications in a configuration file.
Status db file is corrupt. The Status Database has been damaged. Shut down SharePlex and remove the statusdb file, which resides in the data subdirectory of the SharePlex variable-data directory. SharePlex will create another one when you start sp_cop again.
Parameter does not exist in database.

You tried to set a parameter, and you entered the wrong name or the parameter is deprecated for your SharePlex version.

Use the list param command to view the SharePlex parameters for your version and to verify the spelling.

Parameter type checking failed - look in param - defaults file. You might have entered the wrong data type for the parameter. Use the list param command to determine the valid data type.

Unknown service specified.

or...

No such module.

or...

Service may be only one of: post, read, import, export, capture, all.

Valid service (process) names are capture, read, export, import, post. Issue the command again with the correct name.

Command was called with an invalid argument.

or…

Unknown keyword used in command.

The command contains invalid input. Issue the help command to view valid input for the command.
Permission denied for command - check your authorization level. You are not a member of the user group that can issue this command. Issue the authlevel command to view your authorization level.
Default host is not defined: use the ‘host’ command or [on host] option. SharePlex cannot to determine which system you want the command to affect. Either establish a default host with the host command or use the [on host] option with the command that you want to issue (if available).

How to resynchronize source and target objects

The following instructions help you decide how to resynchronize out-of-sync tables.

Manually patch out-of-sync tables

Valid for: All database types

If the number of synchronization errors is small, you can try to repair out-of-sync tables manually. When the Post process detects an out-of-sync condition, it ignores the error and continues to apply the next operations in the post queue. However, Post logs source SQL statements that cause out-of-sync errors to an error file calleID_errlog.sql. (ID is the identifier that SharePlex uses for the target instance, such as the ORACLE_SID or the database name.) You can apply those SQL statements to a target table through the native SQL interface of the database. Because this procedure bypasses the comparison made by Post, the operations should succeed assuming the structure of the target table did not change.

SharePlex stores ID_errlog.sql in the data sub-directory of the variable-data directory on the target system. The entries in the file are similar to the following example:

-- Host (irvlabua) Sid (al920u64)

-- session 2, 1 error --

--

-- [1] Tue Dec 11 13:31:32 2007

-- redolog seq#/offset 26622/26980368

-- redolog timestamp 641050290 (12/11/15 13:31:30)

-- original rowid AAE0m8AAWAAAAFEAAA

-- -- NOT FOUND

delete from “SP_5”.”QA_LOB_DISABLE_INROW” t where rownum = 1 and “KEY”='01';

To apply the SQL manually

  1. Stop user access to the affected source table.
  2. On the target system, open the ID_errlog.sql file.
  3. Apply the SQL statement(s) to the target table.
  4. Reactivate the configuration if you had to make any changes to it.

    sp_ctrl> activate config filename

  5. Allow user access to the source table.

Resynchronize by copying the source tables

Valid for: All database types

This procedure restores synchronization to out-of-sync target tables by applying a copy of the source tables. You only need to resynchronize the tables that are out of synchronization, so users can continue accessing all other tables.

Important! Before you start, review this procedure and see the SharePlex Reference Guide for more information about the commands that are used.

  1. On the source and target systems, make sure sp_cop is running.
  2. On the target system, run sp_ctrl.
  3. [If necessary] On the target system, issue the show sync command to identify the tables that are out of synchronization.

    sp_ctrl> show sync

  4. On the source system, stop activity for the out-of-sync tables.
  5. On the source system, run sp_ctrl.
  6. On the source system, issue the flush command. Note: This command has additional options for use with named queues or multiple targets. See the SharePlex Reference Guide for more information about this command.

    sp_ctrl> flush datasource

  7. On the source system, copy the tables.
  8. On the source system, reactivate the configuration file if you had to make any changes.

    sp_ctrl> activate config filename

  9. On the source system, allow users back onto the source tables.
  10. On the target system, issue the status command until it shows that Post stopped.

    sp_ctrl> status

  11. On the target system, restore the tables.

  12. On the target system, disable or modify triggers, referential integrity constraints and check constraints according to the requirements of your replication strategy.
  13. On the target system, determine the status ID of each message by viewing the Status Database.

    sp_ctrl> show statusdb detail

  14. On the target system, clear each message with the following command.

    sp_ctrl> clear status statusID

  15. On the target system, start the Post process.

    sp_ctrl> start post

Resynchronize with Oracle transportable tablespace

Valid for: Oracle database

The transportable tablespace feature enables you to resynchronize numerous out-of-sync tables quickly and with minimal downtime. To use the transportable tablespace feature, follow the instructions in the Oracle documentation for generating a tablespace set, moving the tablespace set to the target database, and plugging the set into the database. The following instructions contain steps only for using this feature to resynchronize data. It assumes familiarity with using the transportable tablespace feature.

Important! Before you start, review this procedure and see the SharePlex Reference Guide for more information about the commands that are used.

  1. On the source system, set the source tablespace to READ ONLY.

    SQL> ALTER TABLESPACE name READ ONLY;

  2. On the source system, run sp_ctrl.
  3. On the source system, issue the flush command in sp_ctrl. Note: This command has additional options for use with named queues or multiple targets. See the SharePlex Reference Guide for more information.

    sp_ctrl> flush datasource

  4. Export the metadata to an export file according to the Oracle documentation.
  5. When the export is finished, copy the datafiles to a secondary location on the source system. This minimizes the impact on the source database of copying the files to the target system.
  6. On the source system, set the source tablespace(s) to READ WRITE mode.

    SQL> ALTER TABLESPACE name READ WRITE;

  7. On the target system, drop the existing datafiles and tablespaces from the target database so that the copied files can be applied.
  8. Copy the files from the secondary location on the source system to the target system.
  9. On the target system, use the Oracle import utility to import the metadata and the tablespace definitions.
  10. On the target system, set the tablespace(s) to READ WRITE mode.

    SQL> ALTER TABLESPACE name READ WRITE;

    Note: SharePlex must be the only user permitted to have write access to the target tables, unless you are using peer-to-peer replication.

  11. On the source system, reactivate the configuration file if you had to make any changes to it.

    sp_ctrl> activate config filename

  12. On the target system, run sp_ctrl.
  13. On the target system, start the Post process.

    sp_ctrl> start post

Resynchronize with an Oracle hot backup on an active database

Valid for: Oracle database

When you use an Oracle hot backup and the reconcile command to resynchronize a target instance, users can continue to access the production data while the backup is made and applied.

Important:
  • To resynchronize centralized reporting, such as a data warehouse, you cannot use a hot backup from all source systems. One backup would override the data from the previous one. You can use a hot backup of one of the source instances to establish the target instance, and then use another method such as export/import or transportable tablespaces to copy the tables from the other instances.
  • To resynchronize peer-to-peer replication, you must quiet all of the secondary source systems for the duration of this procedure. Move all users to the primary system, and then follow the procedure. After the procedure has been performed on all of the secondary systems, users can resume activity on them.
  • Before you start, review this procedure and see the SharePlex Reference Guide for more information about the commands that are used.

To resynchronize with a hot backup

  1. On the source and target systems, run sp_ctrl.
  2. On the target system, stop the Post process. This allows the replicated data to accumulate in the post queue until the target instance has been recovered and reconciled.

    sp_ctrl> stop post

  3. On the source system, run the Oracle hot backup.
  4. On the source and target systems, verify that sp_cop, sp_ctrl and all SharePlex processes (Capture, Read, Export, Import, Post) are running.

    sp_ctrl> status

  5. Switch log files on the source system.

    • To recover the database to a sequence number, make a note of the highest archive-log sequence number.

    • To recover the database to an Oracle System Change Number (SCN), pick an SCN to recover to on the target database.
  6. Recover the target database from the hot backup:

    • If recovering to a sequence number, recover the database from the hot backup using the UNTIL CANCEL option in the RECOVER clause, and cancel the recovery after Oracle has fully applied the log from the previous step.
    • If recovering to a SCN, recover the database from the hot backup using the UNTIL CHANGE SCN option in the RECOVER clause, and cancel the recovery after Oracle has applied the logs matching the SCN from the previous step.
  7. Open the database with the RESETLOGS option.

  8. On the target system, issue the reconcile command. If you are using named post queues, issue the command for each one. Issue the qstatus command if you are unsure of the queue name.

    • If recovering to a sequence number, substitute the sequence number of the log that you noted in step 5.

      sp_ctrl> reconcile queue queuename for datasource-datadest seq sequence_number

      Example: reconcile queue SysA for o.oraA-o.oraA seq 1234

    • If recovering to a SCN, substitute the SCN that you noted in step 5.

      sp_ctrl> reconcile queue queuename for datasource-datadest scn scn_number

      Example: reconcile queue SysA for o.oraA-o.oraA scn 0123456789

    The reconcile process retains control of sp_ctrl until it is finished, and then the sp_ctrl prompt returns.

  9. On the target system, log onto SQL*Plus as the Oracle user for SharePlex, and run the cleanup.sql script located in the bin sub-directory of the SharePlex product directory. This script truncates and updates the SharePlex tables, which are owned by the SharePlex user. If you are running multiple instances of sp_cop with multiple variable-data directories, there is a SharePlex Oracle user for each one. Make sure you run this script as the SharePlex user that owns the tables you want to restore. The script prompts you for the SharePlex user name and password.

    SQL> @/productdir/bin/cleanup.sql

  10. On the target system, disable or modify the following according to your replication strategy:

    • triggers
    • foreign key constraints
    • cascading delete constraints (or configure SharePlex to ignore them)
    • check constraints
    • scheduled jobs that perform DML
  11. On the source system, reactivate the configuration file if you had to make any changes to it.

    sp_ctrl> activate config filename

  12. On the target system, start the Post process. The two instances are now in synchronization, and SharePlex will continue replicating.

    sp_ctrl> start post

How to restore Oracle archive logs

If you decide to restore the archive logs to enable SharePlex to resume capture and replication, use the following procedure to determine the required archive logs.

  1. Determine the sequence number that Capture needs to resume processing from. Capture stops when it encounters a log wrap and prints a message to the Event Log (event_log) containing the redo log sequence number it needs. You also can find out this number by querying the SHAREPLEX_ACTID table and looking at the SEQNO column, as shown in the following example:

    SQL> select * from splex.shareplex_actid;

    ACTID SEQNO OFFSET AB_FLAG QUE_SEQ_NO_1 QUE_SEQ_NO_2 COMMAND
    ----- ------ -------- -------- ------------- -------------- ------------
    14 114 9757200 0 672101000 0  
  2. Query the Oracle V$LOG_HISTORY table to find out when that sequence number was archived, then copy the logs from that point forward to the source system.

    SQL> select * from V$LOG_HISTORY;

    RECID STAMP THREAD# SEQUENCE# FIRST_CHANGE# FIRST_TIM NEXT_CHANGE#
    ----- ------ -------- -------- ------------- --------------
    111 402941650 1 111 2729501 14-JUL-00 2729548
    112 402941737 1 112 2729548 14-JUL-00 2729633
    113 402941930 1 113 2729633 14-JUL-00 2781791
    114 402942019 1 114 2781791 14-JUL-00 2836155
    115 402942106 1 115 2836155 14-JUL-00 2890539
Related Documents

The document was helpful.

Select Rating

I easily found the information I needed.

Select Rating