This section describes online backup of the Management Server.
When using online backup of the Management Server, the PostgreSQL Point-In-Time Recovery (PITR) mechanism is used in the backup of the metering database.
When using PITR backup and restoration, the following two resource types are backed up and restored:
Base backup
The entire database cluster (file group in which database data is recorded) is backed up.
WAL file
Write-Ahead Log (WAL) files are files in which the contents of updates to the database are recorded.
Backup of the two preceding resource types is taken when the commands for the two corresponding backup methods (base backup and differential backup) are executed.
Resources that are backed up and restored using PITR are stored in the directory that has been specified in the settings file, regardless of the storage destination specified when the command was executed. Refer to "8.5 Online Backup Settings for Metering" for information on how to set this.
Each of the backup methods for performing online backup of the Management Server is described below.
Base backup
Base backup is the backup of the entire database cluster (file group in which database data is recorded).
Execute the command shown below.
Refer to "1.8.11 rcxmgrbackup" of the "Reference Guide (Resource Management) CE" for information on the command.
>Installation_folder\SVROR\Manager\bin\rcxmgrbackup -dir directory -base [-cleanup] <RETURN>
# /opt/FJSVrcvmr/bin/rcxmgrbackup -dir directory -base [-cleanup] <RETURN>
Note
The backup command may not end normally at times, so do not perform the following operations:
Forced end using Ctrl+C during execution of the backup command
Stopping the Manager during execution of the backup command
If the operations listed above have been performed, the following action will be required, depending on the status:
Execute the command shown below the next time and any subsequent times the base backup does not end normally.
When this command is executed, base backup will end normally.
Refer to Section "2.9 ctmg_resetbackuperror (Reset Base Backup Error)" of the "Reference Guide CE" for information on the command.
<Installation_folder>\RCXCTMG\bin\ctmg_resetbackuperror.bat
The processing result is output as standard output.
The contents and meaning of the processing result are shown in the table below.
Processing Result | Return Value | Message |
|---|---|---|
Normal end | 0 | Successfully reset the base-backup error. |
Error | Other than 0 | Failed to reset the base-backup error. |
If starting of the Manager fails
When operations continue after a while without the command mentioned above (ctmg_resetbackuperror) being executed, and then the Manager is stopped, subsequent starts of the Manager may fail. If this happens, an error message will be output to the database log files, as follows:
Database log files
Installation_folder\RCXCTMG\Charging\log\psql-nn.log (*1)
*1) The "nn" part is a 2-digit numeral indicating the day on which the log was output.
Error message
Example: If the access control database failed to start
LOG: could not open file "pg_xlog/xxxxxxxx" (log file 0, segment xx): No such file or directory (*2) LOG: invalid checkpoint record PANIC: could not locate required checkpoint record HINT: If you are not restoring from a backup, try removing the file "C:/Fujitsu/ROR/RCXCTMG/Charging/pgsql/data/backup_label".
*2) The "xxxxxxxx" and "xx" parts of the log are undefined.
In a case like this, delete the file shown below. When this file is deleted, start of the Manager will end normally. When this file is deleted, start of the Manager will end normally.
Installation_folder\RCXCTMG\Charging\pgsql\data\backup_label
Differential backup
With differential backup, the contents of updates to the database are output to multiple files in 16 MB-sized lots.
These files are called Write-Ahead Logging (WAL) files.
Usually, for each 16 MB written, the WAL file being written to is switched, and the WAL file for which writing has completed is saved to the "wal" directory under the backup directory.
Execute the rcxmgrbackup command.
Periodically executing the rcxmgrbackup command allows the contents of updates to the database that are recorded in saved WAL files to be maintained for a certain period of time. For example, when the rcxmgrbackup command is executed every hour, the contents of updates performed in the most recent one-hour period will be saved.
Refer to "1.8.11 rcxmgrbackup" of the "Reference Guide (Resource Management) CE" for information on the command.
>Installation_folder\SVROR\Manager\bin\rcxmgrbackup -dir directory [-cleanup] <RETURN>
# /opt/FJSVrcvmr/bin/rcxmgrbackup -dir directory [-cleanup] <RETURN>
