The following diagram illustrates the CUBRID database volume structure. As you can see, the database is divided into three volumes: permanent, temporary and backup. This chapter will examine each volume and its characteristics.
Permanent volume is a database volume that exists permanently once it is created. Its types include generic, data, temp, index, control, active log and archive log.
For efficient management, the volume type to be added to the database can be specified as one of the followings: data, temp or index. If data usage is not specified, it is specified as a generic volume.
Data volume is a space for storing data such as instances, tables and multimedia data.
Temp volume is a space where the intermediate and final results of query processing and sorting are temporarily stored. It is also called the permanent temp volume to distinguish it from the temporary temp volume that will be described below. The temp volume is a permanently-occupied space, meaning that data in the space is temporary stored and later destroyed. Therefore, when CUBRID restarts, the data in the temp volume space is initialized and the related log information is removed.
The example of query that can use permanent temp volume or temporary temp volume is as follows:
When executing the queries above, the temp volume is used after exhausting the memory space (the space size is determined by the system parameter temp_file_memory_size_in_pages specified in cubrid.conf) assigned to store SELECT results or sort the data. The order in which the storage space is used to store the results of query processing and sorting is as follows: When the current storage space is exhausted, the next storage space is used.
temp_file_memory_size_in_pages memory secured by the system parameter
Index volume is a space that holds the index information for query processing or integrity constraint checks.
The control file contains the volume, backup and log information in the database.
Control files include the information about locations of database volumes, backups and logs. Since these files will be read when the database restarts, users must not modify them arbitrarily.
Active log is a log that contains recent changes to the database. If a problem occurs, you can use active and archive logs to restore the database completely up to the point of the last commit before the occurrence of the fault.
Archive log is a volume to store logs continuously created after exhausting available active log space that contains recent changes. The archive log volume will be generated only after exhausting available active log volume space. Because archive log is not automatically destroyed, it is required to configure archive log to be deleted by system.
Note To get information on the conditions in which archive log can be deleted, see Managing Archive Log.
Background Archive Log
Background archive log is a volume used in the background with log archiving temporarily before creating archive logs. It is created as the same volume size as active log and stored.
Temporary volume has the opposite meaning to the permanent volume. That is, the temporary volume is a storage created only when the accumulated data exceeds the space specified by the user as the permanent volume. The temporary volume is destroyed when the server process terminates. One of such volumes created or destroyed temporarily is the temporary temp volume.
Temporary Temp Volume
Temporary temp volume is a temporary volume created temporarily by the system after exhausting the space specified as the permanent temp volume, whereas the temporary volume belongs to the permanent volume with the permanent space specified. Because cost to create a temporary temp volume is high, DBA should assign an appropriate size of volume depending on situation to enhance performance.
DBA should consider space where temporary temp volume can be created when creating a database. Once temporary temp volume is created, it is maintained until a database restarts and its size cannot be reduced. It is recommended to make temporary temp volume automatically delete by restarting a database if its size is too big.
Backup volume is a database snapshot; based on such backup and log volumes, you can restore transactions to a certain point of time.
You can use the cubrid backupdb utility to copy all the data needed for database restore, or configure the backup_volume_max_size_bytes parameter value in the database configuration file (cubrid.conf) to adjust the backup volume partitioning size.