Point
It is recommended that you back up the following resources before starting the upgrade.
By preparing a backup, in cases when an error occurs during the upgrade, it is possible to return the system to its state prior to the upgrade, simply by recovering (restoring) the system from the backup.
System (the primary node and the secondary node)
Shared disk for shared data of Management Server transactions
Shared disk for shared data of Managed Server transactions
Login to the server on which to perform the tasks. Perform operations after logging in with root (superuser).
Ensure that there is enough available disk space (650 MB in a Solaris environment, 200 MB in a Linux environment) on the server where the upgrade is performed on the secondary node.
In cases when upgrade installation is executed in a state where the capacity required for the upgrade is not available, the following message is displayed and the installation terminates in an error. Re-execute the upgrade after securing the capacity required for the installation.
[Output Message (in cases when the directory of the destination of the installation is "/opt")]
ERROR:Disk /opt has an insufficient free space. Please execute it again after increasing the disk area.
On the secondary node, check whether the cluster application (transaction) to which ETERNUS SF Manager belongs has stopped.
If not stopped, refer to the relevant cluster software manuals for information on stopping it.
Ensure that there is enough available disk space (650 MB in a Solaris environment, 200 MB in a Linux environment) on the server where the upgrade is performed on the primary node.
In cases when upgrade installation is executed in a state where the capacity required for the upgrade is not available, the following message is displayed and the installation terminates in an error. Re-execute the upgrade after securing the capacity required for the installation.
[Output Message (in cases when the directory of the destination of the installation is "/opt")]
ERROR:Disk /opt has an insufficient free space. Please execute it again after increasing the disk area.
On the primary node, stop the cluster application (transaction) to which ETERNUS SF Manager belongs. Refer to the relevant cluster software manuals for information on stopping it.
However, the shared disk for shared data of ETERNUS SF Manager must be mounted.
If Managed Server transactions coexist in a clustered system, perform the following.
On the secondary node for the target transactions, check that the Managed Server transactions have stopped.
If not stopped, stop them on the secondary node.
If multiple Managed Server transactions exist, perform this procedure for each secondary node for Managed Server transactions.
Refer to the relevant cluster software manuals for information on stopping the Managed Server transactions.
On the primary node for the target transactions, stop the Managed Server transactions.
Refer to the relevant cluster software manuals to stop the Managed Server transactions.
However, the shared disk for shared data of Managed Server transactions must be mounted.
If multiple Managed Server transactions exist, perform this procedure for each primary node for Managed Server transactions.
Stop the local transactions on all the nodes.
Stop the communication daemon for the local transactions of the AdvancedCopy Manager's manager.
Refer to "Starting and Stopping Communication Daemon" in the AdvancedCopy Manager Operation Guide relevant to the OS of the Management Server for this version for information on stopping the daemon.
On the primary node and the secondary node, save the start and stop script of the Management Server transaction.
In cases when Managed Server transactions coexist in the cluster system, move the AdvancedCopy Manager daemon start and stop script of the Managed Server transactions to a backup location.
On the primary node and the secondary node, move the AdvancedCopy Manager daemon start and stop script of the Managed Server transactions to a backup location.
Refer to "C.4 Procedures for Upgrade Installation of Manager Feature of Express / Storage Cruiser / AdvancedCopy Manager Version 16.x (for Solaris, Linux)" and perform the Upgrade Installation.
Setup the starting process of ETERNUS SF Manager to be controlled by cluster software instead of operating system.
For Solaris environments
Execute the following command.
# ls /etc/rc*.d/*SFmanager*
Rename all of the files output by executing the command above.
Rename by adding "X" to the start of the file names.
If a file name already exists, delete the existing file before renaming.
Example
The command execution example is as follows.
# mv /etc/rc2.d/S99startSFmanager /etc/rc2.d/XS99startSFmanager
For Linux environments
For Red Hat Enterprise Linux 8 or Red Hat Enterprise Linux 7 environments
Execute the following commands.
# /usr/bin/systemctl disable fjsvesfcm-internalDB.service # /usr/bin/systemctl disable fjsvesfcm-webserver.service # /usr/bin/systemctl disable fjsvesfcm-system.service # /usr/bin/systemctl disable fjsvssast.service
For Red Hat Enterprise Linux 6 environments
Execute the following commands to rename files.
If a file name already exists, delete the existing file before renaming.
# mv /etc/rc0.d/K00startSFmanager /etc/rc0.d/XK00startSFmanager # mv /etc/rc1.d/K00startSFmanager /etc/rc1.d/XK00startSFmanager # mv /etc/rc2.d/K00startSFmanager /etc/rc2.d/XK00startSFmanager # mv /etc/rc3.d/S99startSFmanager /etc/rc3.d/XS99startSFmanager # mv /etc/rc4.d/S99startSFmanager /etc/rc4.d/XS99startSFmanager # mv /etc/rc5.d/S99startSFmanager /etc/rc5.d/XS99startSFmanager # mv /etc/rc6.d/K00startSFmanager /etc/rc6.d/XK00startSFmanager
Unmount the shared disk for shared data of ETERNUS SF Manager.
Point
If the Upgrade Installation has ended in an error, after removing the cause of the failure of the Upgrade Installation, re-execute the installation from the execution of the installation shell.
Mount the shared disk for shared data of ETERNUS SF Manager.
Refer to "C.4 Procedures for Upgrade Installation of Manager Feature of Express / Storage Cruiser / AdvancedCopy Manager Version 16.x (for Solaris, Linux)" and perform the Upgrade Installation.
Setup the starting process of ETERNUS SF Manager to be controlled by cluster software instead of operating system.
For Solaris environments
Execute the following command.
# ls /etc/rc*.d/*SFmanager*
Rename all of the files output by executing the command above.
Rename by adding "X" to the start of the file names.
If a file name already exists, delete the existing file before renaming.
Example
The command execution example is as follows.
# mv /etc/rc2.d/S99startSFmanager /etc/rc2.d/XS99startSFmanager
For Linux environments
For Red Hat Enterprise Linux 8 or Red Hat Enterprise Linux 7 environments
Execute the following commands.
# /usr/bin/systemctl disable fjsvesfcm-internalDB.service # /usr/bin/systemctl disable fjsvesfcm-webserver.service # /usr/bin/systemctl disable fjsvesfcm-system.service # /usr/bin/systemctl disable fjsvssast.service
For Red Hat Enterprise Linux 6 environments
Execute the following commands to rename files.
If a file name already exists, delete the existing file before renaming.
# mv /etc/rc0.d/K00startSFmanager /etc/rc0.d/XK00startSFmanager # mv /etc/rc1.d/K00startSFmanager /etc/rc1.d/XK00startSFmanager # mv /etc/rc2.d/K00startSFmanager /etc/rc2.d/XK00startSFmanager # mv /etc/rc3.d/S99startSFmanager /etc/rc3.d/XS99startSFmanager # mv /etc/rc4.d/S99startSFmanager /etc/rc4.d/XS99startSFmanager # mv /etc/rc5.d/S99startSFmanager /etc/rc5.d/XS99startSFmanager # mv /etc/rc6.d/K00startSFmanager /etc/rc6.d/XK00startSFmanager
Unmount the shared disk for shared data of ETERNUS SF Manager.
Point
If the Upgrade Installation has ended in an error, after removing the cause of the failure of the Upgrade Installation, re-execute the installation from the execution of the installation shell.
The tasks below need to be performed on the primary node.
The shared disk for shared data of ETERNUS SF Manager must be mounted.
In the version level of this ETERNUS SF Manager, Symfoware is not used.
To upgrade from Version 16.1 or later, the procedures listed below are not needed.
To upgrade from Version 16.0 or earlier, follow the procedures listed below to uninstall the Symfoware Server.
Confirm that other products are not utilizing the Symfoware. If Symfoware is in use by another product, refrain from uninstalling Symfoware.
Start the uninstaller of Symfoware to delete the Symfoware package.
# /opt/symfoware/setup/symfo_remove
Confirmation message to uninstall Symfoware is displayed.
To perform the uninstallation, enter "y". To cancel the process, enter "q".
When "y" is entered, a message allowing to select the function to uninstall is shown. Input "all" or "number to delete" to select the function to uninstall.
A message to re-confirm uninstallation appears. Enter "y" to perform the uninstallation and "q" when it is not required.
Start checking the installation environment. Check of installation environment is ended. WARNING: There is a possibility that another product uses "Symfoware Server Enterprise Edition". Continue the uninstallation? y: Continue the uninstallation q: End directly without uninstallation [y,q]: y Uninstallation of "Symfoware Server Enterprise Edition" begins. Please select the functions to be uninstalled. 1: Base function (Server function, Client function) Separate multiple selections with ",". (Example: 1, 2) [all:All functions,1,q]: all Selected functions - Base function (Server function, Client function) Start the uninstallation of the above functions? y: Uninstallation begins n: Select the functions again q: End directly without uninstallation [y,n,q]: y Uninstallation begins.
The following message is displayed after the Symfoware uninstallation is completed normally.
Uninstallation of "Symfoware Server Enterprise Edition" is ended normally.
Edit the environment setting files for ETERNUS SF Manager.
Edit the following files:
<Mount point of shared disk for shared data of ETERNUS SF Manager>/etc/opt/swstorage/clsetup.ini
<Mount point of shared disk for shared data of ETERNUS SF Manager>/etc/opt/swstorage/swstg.ini
Change the version information within each file, as follows.
Descriptive Contents of Version Information | |
|---|---|
Before | After |
Version=V16.x | Version=V16.9 |
Note
Do not change anything other than the version information.
If Managed Server transactions coexist in the clustered system, edit the environment setting files for AdvancedCopy Manager on the shared disk for shared data of Managed Server transactions, on the primary node for a target service.
On the primary node for target transactions, edit the following files:
<Mount point of shared disk for shared data of Managed Server transaction>/etc/opt/swstorage/clsetup.ini
<Mount point of shared disk for shared data of Managed Server transaction>/etc/opt/swstorage/swstg.ini
Change the version information within each file, as follows.
Descriptive Contents of Version Information | |
|---|---|
Before | After |
Version=V16.x | Version=V16.9 |
Note
Do not change anything other than the version information.
Recover the activation suspension script of Management Server tasks saved in "7.2.2.1 Preparing for Upgrade".
In cases when Managed Server transaction coexist in the cluster system, recover the AdvancedCopy Manager daemon start and stop script of the Managed Server transaction saved in "7.2.2.1 Preparing for Upgrade".
Perform the kernel parameter tuning.
See
Refer to "Before Installation" in "Installation of ETERNUS SF Manager" in the Installation and Setup Guide for this version for information on tuning.
The tasks below need to be performed on the secondary node.
In the version level of this ETERNUS SF Manager, Symfoware is not used.
To upgrade from Version 16.1 or later, the procedures listed below are not needed.
To upgrade from Version 16.0 or earlier, follow the procedures listed below to uninstall the Symfoware Server.
Confirm that other products are not utilizing the Symfoware. If Symfoware is in use by another product, refrain from uninstalling Symfoware.
Start the uninstaller of Symfoware to delete the Symfoware package.
# /opt/symfoware/setup/symfo_remove
Confirmation message to uninstall Symfoware is displayed.
To perform the uninstallation, enter "y". To cancel the process, enter "q".
When "y" is entered, a message allowing to select the function to uninstall is shown. Input "all" or "number to delete" to select the function to uninstall.
A message to re-confirm uninstallation appears. Enter "y" to perform the uninstallation and "q" when it is not required.
Start checking the installation environment. Check of installation environment is ended. WARNING: There is a possibility that another product uses "Symfoware Server Enterprise Edition". Continue the uninstallation? y: Continue the uninstallation q: End directly without uninstallation [y,q]: y Uninstallation of "Symfoware Server Enterprise Edition" begins. Please select the functions to be uninstalled. 1: Base function (Server function, Client function) Separate multiple selections with ",". (Example: 1, 2) [all:All functions,1,q]: all Selected functions - Base function (Server function, Client function) Start the uninstallation of the above functions? y: Uninstallation begins n: Select the functions again q: End directly without uninstallation [y,n,q]: y Uninstallation begins.
The following message is displayed after the Symfoware uninstallation is completed normally.
Uninstallation of "Symfoware Server Enterprise Edition" is ended normally.
Recover the activation suspension script of Management Server tasks saved in "7.2.2.1 Preparing for Upgrade".
In cases when Managed Server transaction coexist in the cluster system, recover the AdvancedCopy Manager daemon start and stop script of the Managed Server transaction saved in "7.2.2.1 Preparing for Upgrade".
Perform the kernel parameter tuning.
See
Refer to "Before Installation" in "Installation of ETERNUS SF Manager" in the Installation and Setup Guide for this version for information on tuning.
Check that the cluster application (transaction) has stopped.
If not stopped, stop the cluster application (transaction) to which ETERNUS SF Manager belongs.
Refer to the relevant cluster software manuals for information on stopping it (transaction).
After performing "7.2.2.4 Tasks to Be Performed After Upgrade (Primary Node)" and "7.1.2.5 Tasks to Be Performed After Upgrade (Secondary Node) ", perform the following procedure:
Check that the cluster application (transaction) has started on the primary node.
If not started, start the cluster application (transaction) to which ETERNUS SF Manager belongs.
Refer to the relevant cluster software manuals for information on starting it.
If Managed Server transactions coexist in a clustered system, start the Managed Server transactions on the primary node for the target transactions.
Refer to the relevant cluster software manuals to start the Managed Server transactions.
If multiple Managed Server transactions exist, perform this procedure for each Managed Server transaction.
Start the local transactions on all the nodes.
Start the communication daemon for local transactions of AdvancedCopy Manager's manager.
Refer to "Starting and Stopping Communication Daemon" in the AdvancedCopy Manager Operation Guide for the operating system of the Management Server for this version for information on starting the daemon.
Start the SNMP Trap monitoring daemon on all local nodes. No action is necessary if it is already started. Refer to the following section in the Storage Cruiser Operation Guide for this version for information on how to start the daemon.
For Solaris environments
"Starting and Stopping Daemons (Solaris Environment)"
For Linux environments (Only if the ETERNUS SF SNMP Trap Monitoring Daemon was used with the previous version)
"Starting and Stopping Daemons (Linux Environment)"
Subsequently, go to "7.2.3 Resuming Operation".
Point
Even if the message "A1001 : Could not establish connection with SNMP Trap daemon/service" is output to the system log when the cluster application (transaction) is started, no action is necessary.
