Hitachi

Hitachi Advanced Database Setup and Operation Guide

8.6.4 Steps to take when version upgrading fails

This section explains the steps to take when upgrading of the HADB server version fails.

Steps to take

  1. Check the messages.

    Use a text editor to open the server message log file and check the error messages that have been output. The server message log file is /spool/adbmessageXX.log in the server directory. XX in the file name is a sequential number between 01 and 04. Open the file with the latest modification date and time.

    If none of the messages listed below have been output, proceed to step 2.

    If any of the following messages have been output, take the actions indicated in the referenced subsections of this chapter:

  2. Take action according to the Action section of the relevant message.

    Based on the ID of the output error message, see the relevant page in the manual HADB Messages. Take the step described in the Action section for that message.

    If the Action section says, Contact the customer support center, proceed to step 3.

  3. Execute the adbinfoget command.

    If the Action section says, Contact the customer support center, you need to acquire troubleshooting information to investigate the cause of the error.

    Execute the adbinfoget command as explained in 14.2 Collecting troubleshooting information (adbinfoget command). When you do so, specify the -f option for the command to acquire the error information (core files) as well. For details about the -f option, see adbinfoget (Collect Troubleshooting Information) in the manual HADB Command Reference.

    When you contact the customer support center, send along the troubleshooting information you have acquired.

After you have taken whichever action was indicated as the response to the message that was output, restart the HADB server. If an attempt to restart the HADB server results in an error and version upgrading fails, you will have to recover the database from the backup. See (5) Recovering the database from a backup.

Organization of this subsection

(1) Steps to take when the KFAA30811-E message is output

In this case, version upgrading failed because the number of schemas defined exceeds 30,000. You must first restore the HADB server to the previous version, and then take the steps necessary to recover from the error.

To restore the HADB server to the previous version and recover from the error:

Procedure

  1. Delete the server directory.

    Delete the server directory by following the steps in (1) Deleting the server directory under 8.12.1 Uninstallation procedure.

  2. Recover the previous version of the server directory from a backup.

    From a backup acquired earlier, copy the previous version of the server directory and return it to the storage location from which it was deleted in step 1.

  3. Start the HADB server.

    Use the adbstart command to start the HADB server.

  4. Reduce the number of schemas to 30,000 or fewer.

    Use the DROP SCHEMA statement to deleted unnecessary schemas so that the number of defined schemas is 30,000 or fewer. Version upgrading cannot be executed if the number of schemas exceeds 30,000.

  5. Terminate the HADB server.

    Use the adbstop command to terminate the HADB server.

  6. Confirm that the HADB server terminated normally.

    Check whether the HADB server terminated normally by following the steps in (12) Terminating the HADB server normally under 8.6.1 Steps to take before upgrading the server version.

After you have confirmed that the HADB server terminated normally, perform the version upgrade again by following the steps in 8.6.2 Upgrading the server version.

To Page Top

(2) Steps to take when the KFAA41210-E message is output

In this case, version upgrading failed because the number of user log files was insufficient. You must first restore the HADB server to the previous version, and then take the steps necessary to recover from the error.

To restore the HADB server to the previous version and recover from the error:

Procedure

  1. Delete the server directory.

    Delete the server directory by following the steps in (1) Deleting the server directory under 8.12.1 Uninstallation procedure.

  2. Recover the previous version of the server directory from a backup.

    From a backup acquired earlier, copy the previous version of the server directory and return it to the storage location from which it was deleted in step 1.

  3. Check the value specified for the adb_log_usrfile_num operand in the server definition.

    Check whether 1 is specified for the adb_log_usrfile_num operand in the server definition. If 1 is specified, change it to at least 2.

  4. Start and terminate the HADB server.

    Use the adbstart command to start the HADB server. Then, use the adbstop command to terminate it.

  5. Confirm that the HADB server terminated normally.

    Check whether the HADB server terminated normally by following the steps in (12) Terminating the HADB server normally under 8.6.1 Steps to take before upgrading the server version.

After you have confirmed that the HADB server terminated normally, perform the version upgrade again by following the steps in 8.6.2 Upgrading the server version.

To Page Top

(3) Steps to take when the KFAA50038-E message is output

The steps to take depend on which of the following was output to variable aa....aa in the KFAA50038-E message:

(a) If variable aa....aa is "of an abnormal termination the last time it ended"

Version upgrading failed because the HADB server terminated abnormally. You must first restore the HADB server to the previous version, and then take the steps necessary to recover from the error.

Multi-node function

If you are using the multi-node function, restore the previous version of the HADB servers on all nodes.

To restore the HADB server to the previous version and recover from the error:

Procedure

  1. Delete the server directory.

    Delete the server directory by following the steps in (1) Deleting the server directory under 8.12.1 Uninstallation procedure.

    Multi-node function

    If you are using the multi-node function, perform this step on all nodes.

  2. Recover the previous version of the server directory from a backup.

    From a backup acquired earlier, copy the previous version of the server directory and return it to the storage location from which it was deleted in step 1.

    Multi-node function

    If you are using the multi-node function, perform this step on all nodes.

  3. Start and terminate the HADB server.

    Use the adbstart command to start the HADB server. Then, use the adbstop command to terminate it.

    Multi-node function

    If you are using the multi-node function, start the HADB servers in the multi-node configuration as explained in 16.4.1 Starting the HADB servers in the multi-node configuration. Then, terminate the HADB servers in the multi-node configuration as explained in 16.4.2 Terminating HADB servers in the multi-node configuration.

  4. Confirm that the HADB server terminated normally.

    Check whether the HADB server terminated normally by following the steps in (12) Terminating the HADB server normally under 8.6.1 Steps to take before upgrading the server version.

After you have confirmed that the HADB server terminated normally, perform the version upgrade again by following the steps in 8.6.2 Upgrading the server version.

(b) If variable aa....aa is "the adbstart command was never been executed in previous versions"

Version upgrading failed because the previous version of the HADB server was never started.

You need to execute the adbinit command to initialize the database. For details about how to initialize the database, see 9.2 Initializing a database (creating data DB areas).

(c) If variable aa....aa is "of using the multiple node facility"

Version upgrading failed because an attempt was made to perform version upgrading together with an attempt to configure the system to use the multi-node function on an HADB server whose version is earlier than 03-00.

Version upgrade processing will fail if you specify operands of the multi-node function in the server definition before you upgrade the version. After you have upgraded your HADB server version, change the settings so that the system can use the multi-node function.

Procedure

  1. Delete the adb_sys_multi_node_info operand from the server definition.

    Delete the adb_sys_multi_node_info operand from the server definition so that the multi-node function is not used.

    Alternatively, comment out the adb_sys_multi_node_info operand in the server definition. Specify a hash mark (#) at the beginning of the line containing the operand to treat that line as a comment.

  2. Start the HADB server.

    Start the HADB server by executing the adbstart command while the multi-node function is not used. Version upgrading starts.

    After the adbstart command has executed, the KFAA91107-Q message is output, asking whether you want to upgrade the database version. Press Y to upgrade.

    If the adbstart command terminates normally, upgrading of the HADB server version has been completed.

  3. Terminate the HADB server normally.

    After the adbstart command has terminated normally and the upgrade processing has been completed, terminate the HADB server normally. To do this, execute the adbstop command.

  4. Specify the adb_sys_multi_node_info operand in the server definition.

    Once the upgrade processing is completed, you can use the multi-node function. Specify the adb_sys_multi_node_info operand in the server definition that was deleted in step 1.

  5. Start the HADB server in the multi-node configuration.

    After you have specified the adb_sys_multi_node_info operand in the server definition, start the HADB server in the multi-node configuration as explained in 16.4.1 Starting the HADB servers in the multi-node configuration.

To Page Top

(4) Steps to take when the KFAA51400-E message is output

Version upgrading failed because the database created by the HADB server version 02-00 did not contain any schema.

You need to execute the adbinit command to initialize the database. For details about how to initialize the database, see 9.2 Initializing a database (creating data DB areas).

To Page Top

(5) Recovering the database from a backup

If an attempt to restart the HADB server results in an error after upgrading of the HADB server has failed and you have taken whichever action was indicated as the response to the message that was output, you will have to recover the database from a backup.

The procedure below explains how to recover the database from a backup and re-configure the previous version of HADB server.

Multi-node function

If you are using the multi-node function, perform this step on all HADB servers on all nodes.

Procedure

  1. Delete the server directory.

    Delete the server directory by following the steps in (1) Deleting the server directory under 8.12.1 Uninstallation procedure.

    Multi-node function

    If you are using the multi-node function, perform this step on all nodes.

  2. Recover the previous version of server directory from a backup.

    From a backup acquired earlier, copy the previous version of the server directory and return it to the storage location from which it was deleted in step 1.

    Multi-node function

    If you are using the multi-node function, perform this step on all nodes.

  3. Recover the DB directory from a backup.

    Recover the DB directory from the backup by following the procedure in 10.3.2 Recovering the database from the backup.

    Multi-node function

    If you are using the multi-node function, recover the DB directory from the backup as described in 16.8.2 Recovering a database from a backup.

  4. Check the value specified for the adb_db_path operand in the server definition.

    Check whether the path name of the DB directory recovered in step 3 is defined in the adb_db_path operand in the server definition.

  5. Start and terminate the HADB server.

    Use the adbstart command to start the HADB server. Then, use the adbstop command to terminate it.

    Multi-node function

    If you are using the multi-node function, start the HADB servers in the multi-node configuration as explained in 16.4.1 Starting the HADB servers in the multi-node configuration. Then, terminate the HADB servers in the multi-node configuration as explained in 16.4.2 Terminating HADB servers in the multi-node configuration.

  6. Confirm that the HADB server terminated normally.

    Check whether the HADB server terminated normally by following the steps in (12) Terminating the HADB server normally under 8.6.1 Steps to take before upgrading the server version.

If the HADB server terminated normally, you can rebuild the previous version of the HADB server. If you need to upgrade the HADB server version, see 8.6.2 Upgrading the server version.

To Page Top

All Rights Reserved. Copyright (C) 2012, 2020, Hitachi, Ltd.