G.4 Notes on performing migration

Note the following when performing migration:

• The version of migration-source PFM - Agent for Virtual Machine must be 09-01 or later. If that version is 09-00 or earlier, update it to 09-01 or later. If the backup file to be migrated was created by 09-00 or an earlier version, convert the data model by executing the jpctool db dmconvert command.

• The Store version of the PFM - Agent for Virtual Machine instance environment must be 2.0. If the Store version is 1.0, update it to 2.0 by executing the jpcconf db vrset command before migration.

• You cannot migrate any backup file created in a PFM - Agent for Virtual Machine instance environment whose Store version is 1.0.

• If the migration-destination PFM - RM for Virtual Machine instance already exists, the VM_Type value of the source instance and the VM_Type value of the destination instance must match.

• If the VM_Host value of the migration-source instance is not 1 through 32 bytes long or has characters other than single-byte alphanumeric characters and hyphens (-), set the Target Host value by specifying the -targethost option.

• The Store database is larger after migration than before migration. Be careful about this especially when you perform migration by adding monitoring targets to an existing PFM - RM for Virtual Machine instance. Migration is impossible if one of the database files that make up the Store database becomes larger than 2,048 MB.

• To perform migration by adding monitoring targets to an existing PFM - RM for Virtual Machine instance, the directory containing the destination Store database must temporarily have free space that is twice the size of the Store database.

• When migration is performed by adding monitoring targets to an existing PFM - RM for Virtual Machine instance, the destination Store database is backed up automatically by the migration command (rather than manually backed up by jpctool db backup). For this reason, migration might take time if the size of the destination Store database is large.

• If you use the backup2store or backup2backup subcommand to perform migration from a backup directory, check the access control settings of the backup directory specified for the -aobkpath option and all files in that directory. Make sure that the access control settings of the directory and files permit reading and writing.

• You can execute multiple instances of the jpcaovm2rmvm command concurrently. In this case, do not specify the same PFM - Agent for Virtual Machine instance as the migration source or the same PFM - RM for Virtual Machine instance as the migration destination in multiple commands.

• While the jpcaovm2rmvm command is being executed, do not add or change the PFM - RM for Virtual Machine instances or monitoring targets.

• While the jpcaovm2rmvm command is being executed, do not use any other commands (such as jpcconf, jpctool, jpcctrl, jpcspm, jpcstart, and jpcstop).

• You cannot execute the jpcaovm2rmvm command while the Store service is being backed up or exported.

• If the jpcaovm2rmvm command is interrupted by, for example, pressing Ctrl + C, invalid instances or monitoring targets might be created. Manually delete any invalid PFM - RM for Virtual Machine instances or monitoring targets that are created.

• If the jpcaovm2rmvm command is interrupted by, for example, pressing Ctrl + C, an unknown value might be returned. Therefore, if the command is interrupted, ignore the returned value.

• In the following cases, because the performance data for the deleted monitoring target remains in the Store database, you cannot perform migration by specifying the same monitoring target name as the deleted one:

  1. After performance data is collected for a certain period from a monitoring target that was created by the jpcconf target setup command, the jpcconf target unsetup command is used to delete the monitoring target.

  2. After the jpcaovm2rmvm command is used to migrate performance data for a monitoring target, the jpcconf target unsetup command is used to delete the monitoring target.

  3. The KAVL20479-E message, which reports a Store database migration error, is output while performance data is being migrated by using the jpcaovm2rmvm command. Alternatively, an invalid monitoring target that was created during migration is deleted by using the jpcconf target unsetup command.

  Take one of the following actions:

  1. Delete the destination instance environment, and then reperform migration, beginning with instance creation.

  2. Rename the monitoring target, and then re-execute the jpcaovm2rmvm command.

  3. Use the defineonly subcommand to migrate only definitions, and then use the jpcconf db import command to import a backup file that has been migrated by using the backup2backup subcommand.

• The default Collection Interval value of PFM - Agent for Virtual Machine is 60 seconds, whereas the default Interval value of the PFM - RM for Virtual Machine collection process is 300 seconds. If there are problems with the collection interval and offset settings, a KAVL20516-W error might be output to the shared message log.

  To avoid this error, make sure that the Interval value does not exceed the Collection Interval value. If the error occurs frequently even when the Interval value does not exceed the Collection Interval value, increase the Collection Interval value or reduce the number of hosts monitored in the instance environment.

  For details about the performance data collection interval and offset settings, see the chapter on designing an operation monitoring system in the JP1/Performance Management Planning and Configuration Guide. For details about the Interval setting for the collection process, see 1.3.1 General procedure for performance data collection tasks.

• Do not redirect the command's standard output. Doing so will output the command execution results including progress information.

• If you perform migration of a cluster environment by using the store2store subcommand, make sure that you specify logical host names for the -aolhost and -rmlhost options. If you use the backup2store or bacup2backup subcommand, make sure that you specify the backup directory for the Store database of PFM - Agent for Virtual Machine. In a cluster environment, after you migrate the environment to the executing node, you must copy the environment definition file from the executing node to the standby node. For details about how to copy the environment definition file, see G.3 Migration in a cluster environment.

• If you want to migrate from the PFM - Agent for Virtual Machine with monitoring VMware with Hyper-Threading enabled, you need to setup and same settings described at 2.1.4(3) Setting PFM - RM for Virtual Machine.

To Page Top