The following overview describes database parameters belonging to the EXTENDED group.
For information about other database parameters, see
To change the values of database parameters, use one of the following database tools: Database Studio, Database Manager CLI, or CCMS (in SAP systems only). More information:
Database Studio, Changing Database Parameters
Database Administration in CCMS, Parameters
Most changes to database parameters take effect only after the database has been restarted.
However, you can change certain database parameters while the database is running, that is in the ONLINE or ADMIN operational state (see the description of individual database parameters for more information). For these changes you can define when they are to become effective:
Until the next restart
After the next restart
Immediately and permanently
Note
Starting with version 7.7.03 of the database software, the names of many database parameters have changed.
Some database parameters have also been designated obsolete. These parameters have either been removed already or will be removed in one of the next database versions.
More information: Obsolete Database Parameters
Name (from version 7.7.03) |
Old Name (up to and including version 7.7.02) |
Description |
Change Restrictions |
More Information |
---|---|---|---|---|
AllowAuthentication |
AUTHENTICATION_ALLOW |
Specifies the authentication methods that are allowed for the database. Possible Values (comma-separated):
If you do not specify a value, then all authentication methods are allowed except those specified using the database parameter DenyAuthentication. |
Can be changed in ONLINE or ADMIN operational state |
|
CAT_CACHE_SUPPLY |
CAT_CACHE_SUPPLY |
Size of the task-specific catalog cache in the working memory that is allocated to user tasks (in pages) For each user task the system reserves a minimum part of the catalog cache (see TaskSpecificCatalogCacheMinSize). Each task may extend its catalog cache, but all task-specific catalog caches together may not exceed the catalog cache size. Note that the catalog cache size influences performance: You achieve the best performance with a catalog cache hit rate of 100%. Use database monitoring to get information about the catalog cache hit rate. |
Changes only apply after a restart of the database. |
Concepts of the Database System, Working Memory Areas |
ClusterWriteThreshold |
CLUSTER_WRITE_THRESHOLD |
Minimum amount of blocks that the database system still stores as clusters in the data area |
Can be changed in ONLINE or ADMIN operational state |
- |
ClusterCompressionFillThreshold |
- |
Maximum number of occupied blocks in clusters eligible for compression If more blocks are occupied in a cluster, the database system does not compress the cluster. |
Can be changed in ONLINE or ADMIN operational state |
- |
ConverterStripes |
CONVERTER_REGIONS |
Number of critical sections into which the converter is divided This division enables parallel writing to the converter. |
The default value is calculated by the database system. You can change the value. Changes only apply after a restart of the database. |
Glossary, Stripe |
ConverterVolumeIdLayout |
VOLUMENO_BIT_COUNT |
Number of bits in the converter block address that are reserved for the logical volume number of a data volume |
We recommend that you do not change the value that was defined when the database was created. If you change the database parameter in the ONLINE or ADMIN operational state, you must restore the database afterwards to update the numbering of the data volume. |
Database Administration, Planning the Database |
DataCacheStripes |
_DATA_CACHE_RGNS |
Number of critical sections into which the data cache is divided This division enables parallel writing to the data cache. |
The default value is calculated by the database system. You can change the value. Changes only apply after a restart of the database. |
|
DateTimeFormat |
DATE_TIME_FORMAT |
System default for the date and time format You can override this date and time format when using database tools (standard_date_mask in Loader) or when using SQL statements (datetimeformat). |
Can be changed in ONLINE or ADMIN operational state |
- |
DeadlockDetectionLevel |
DEADLOCK_DETECTION |
Maximum search level for deadlock detection Any deadlocks that the database system does not detect at this search level are only resolved by a timeout. |
Can be changed in ONLINE or ADMIN operational state |
Parameter RequestTimeout |
DefaultCodePage |
DEFAULT_CODE |
System default for the code attribute, applies only to the column values of the data types CHAR[ACTER], VARCHAR and CLOB |
Can be changed in ONLINE or ADMIN operational state |
SQL Reference Manual, Code Attribute |
DenyAuthentication |
AUTHENTICATION_DENY |
Specifies the authentication methods that are not allowed for this database. Possible Values (comma-separated):
Note that older clients can only connect to the database using the BASIC authentication method. Not allowing this method prevents these clients from being able to connect. |
Can be changed in ONLINE or ADMIN operational state n |
|
DiagnoseHistoryCount |
DIAG_HISTORY_NUM |
Number of history versions in the directory in which the database system stores diagnosis files (backups of the most important log files and memory dumps) after a database error |
Changes only apply after a restart of the database. |
Database Administration, Log Files |
DiagnoseHistoryPath |
DIAG_HISTORY_PATH |
Directory in which the database system stores diagnosis files (backups of the most important log files and memory dumps) after a database error |
Changes only apply after a restart of the database. |
See above |
EnableIndexOnlyStrategy |
OPTIM_INV_ONLY |
Defines whether the SQL optimizer uses the Index Only strategy for joins |
Can be changed in ONLINE or ADMIN operational state |
SQL Optimizer, Search Strategies |
EnableLogBackupToPipe |
LOG_BACKUP_TO_PIPE |
NO: After a log backup has been written to a pipe, the connection to the pipe is closed. The system does not receive any information on whether the log backup was successful. As a result, the log area is not released and cannot be overwritten. Therefore, it may become full. Consequently no changes to data can be made any more. YES: The database system can overwrite the log entries in the log area that have already been written to the pipe. Note that the system does not check whether the log backup to the pipe was successful. |
Changes only apply after a restart of the database. |
Database Administration, Backing Up Log Entries |
EnableQueryRewrite |
OPTIMIZE_QUERYREWRITE |
Prerequisite: Support database parameter QueryAnalysisMode is set to EXTENDED. Specifies whether the system tries to transform SQL statements so that the resulting statement is semantically identical but faster to process. The transformations are purely rule-based and do not take into account any statistical information. The main purpose of transformations is to avoid intermediate result sets. In some cases, it may not be possible to display the statement that results from the transformation as an SQL statement. After an SQL statement has been transformed, the SQL optimizer determines the most cost-effective search strategy for accessing the data. The system can transform the following types of SQL statements: SELECT, CREATE CURSOR FOR SELECT, EXPLAIN SELECT After the system has transformed an SQL statement, the result of the EXPLAIN statement contains the text QUERY REWRITE DONE as well as a list indicating the rules executed and the number how often they were executed. |
Can be changed in ONLINE or ADMIN operational state |
- |
ExclusiveLockRescheduleThreshold |
MAXRGN_REQUEST |
Maximum number of attempts a task should make to access a critical section If this number is exceeded, the task releases its access to the CPU. The CPU can then be accessed by another task belonging to the same user kernel thread. |
Can be changed in ONLINE or ADMIN operational state |
Glossary, Critical Section |
FileDirectorySpinlockPoolSize |
FILEDIR_SPINLOCKPOOL_SIZE |
Number of spinlocks available for reader-writer locks for the internal file directory |
Changes only apply after a restart of the database. |
Glossary, Reader-Writer Lock |
FormatDataVolume |
FORMAT_DATAVOLUME |
Only for test purposes Suppresses the formatting of data volumes when a new database is created |
Changes only apply after a restart of the database. |
Glossary, Volume |
HashJoinSingleTableMemorySize |
MAX_SINGLE_HASHTABLE_SIZE |
Maximum size of the working memory that the database system can use to cache a single table when executing a join (in KB) |
Can be changed in ONLINE or ADMIN operational state |
See above |
HashJoinTotalMemorySize |
MAX_HASHTABLE_MEMORY |
Maximum size of the working memory that the database system can use to cache all tables when executing joins (in KB) |
Can be changed in ONLINE or ADMIN operational state |
Glossary, Join |
HighResolutionTimer |
HIRES_TIMER_TYPE |
Type of time measurement that is used by the kernel for internal operations Use this parameter only for multiprocessor computers on which there could be synchronization problems between the CPUs. |
Changes only apply after a restart of the database. |
- |
HotStandbyDelayTime? |
HS_DELAY_TIME_<NNN> |
Delay in seconds after which the standby database applies the data changes of the master database. The question mark (?) is used as placeholder for the three-digit identification of the standby database. |
Changes only apply after a restart of the database. |
Database Administration, Hot Standby |
HotStandbyNodeName? |
HS_NODE_<NNN> |
Only for hot standby systems Host name or IP address of the database The question mark (?) is used as a placeholder for the three-digit identification of the database. The default value for the master database is HotStandbyNodeName001. |
Changes only apply after a restart of the database. |
See above |
HotStandbyStorageDLLPath |
HS_STORAGE_DLL |
Only for hot standby systems Name of the library through which the system addresses the memory management system The default value is libhsscopy. |
Changes only apply after a restart of the database. |
See above |
HotStandbySyncInterval |
HS_SYNC_INTERVAL |
Only for hot standby systems Minimum time span between two commands of the master database to synchronize the standby databases (in s) |
Changes only apply after a restart of the database. |
See above |
IndexListsMergeThreshold |
OPTIM_MAX_MERGE |
Specifies whether the SQL optimizer uses an index-merge strategy If the number of pages of an index that need to be merged exceeds the value specified in IndexListsMergeThreshold, the SQL optimizer does not use this index for an Index Merge strategy. |
Can be changed in ONLINE or ADMIN operational state |
SQL Optimizer, Search Strategies |
JoinSearchLevel |
JOIN_SEARCH_LEVEL |
Algorithm used by the SQL optimizer for the join sequence search |
Can be changed in ONLINE or ADMIN operational state |
SQL Optimizer, |
JoinSearchTableThreshold4 |
JOIN_MAXTAB_LEVEL4 |
Parameter of the SQL optimizer |
Can be changed in ONLINE or ADMIN operational state |
JoinSearchLevel |
JoinSearchTableThreshold9 |
JOIN_MAXTAB_LEVEL9 |
Parameter of the SQL optimizer |
Can be changed in ONLINE or ADMIN operational state |
JoinSearchLevel |
JoinTableBufferSize |
JOIN_TABLEBUFFER |
Size of the buffer (in KB) per user task for table buffers used during join execution |
Can be changed in ONLINE or ADMIN operational state |
SQL Optimizer, Search Strategies for Joins |
KernelMessageFileSize |
KERNELDIAGSIZE |
Size of the kernel log file (in KB) |
Changes only apply after a restart of the database. |
Database Administration, Log Files |
KernelMessageHistoryCount |
KNLMSG_HISTORY_NUM |
Number of stored old kernel log files At each database start, the system moves the old kernel log file to the log history directory. |
Changes only apply after a restart of the database. |
Database parameter DiagnoseHistoryPath |
MaxKernelMessageArchiveSize |
- |
If this parameter is set to a value other than 0, it defines the maximum size of the KnlMsgArchive file (containing archived messages of the database kernel). If the specified size is reached, the system renames the KnlMsgArchive file to KnlMsgArchive_<timestamp> and creates a KnlMsgArchive file. |
Can be changed in ONLINE or ADMIN operational state |
|
LoadBalancingCheckInterval |
LoadBalancingCheckLevel LOAD_BALANCING_CHK |
Time interval in which the database system checks whether it should use load balancing (in s) |
Can be changed in ONLINE or ADMIN operational state |
Concepts of the Database System, Load Balancing between User Kernel Threads |
LocalRedoLogBufferSize |
LOCAL_REDO_LOG_BUFFER_SIZE |
If you enter a value > 0, then the system allocates a local buffer of this size (in bytes) to each transaction for saving redo log entries. |
Changes only apply after a restart of the database. |
- |
LogQueues |
LOG_QUEUE_COUNT |
Number of log queues If this number is 0, the system calculates the number from the parameter MaxLogQueues |
Changes only apply after a restart of the database. |
Glossary, Log Queue |
LogQueueSize |
LOG_IO_QUEUE |
Size of the individual log queues (in pages) |
Changes only apply after a restart of the database. |
See above |
MaxEvents |
_MAXEVENTS |
Maximum number of events in the working memory that can be processed by the system |
Changes only apply after a restart of the database. |
Glossary, Event |
MaxExclusiveLockCollisionLoops |
MP_RGN_LOOP |
Maximum number of times a task attempts to access a critical section that has been locked by another task If this number is exceeded, the accessing task changes its state to Waiting. |
Can be changed in ONLINE or ADMIN operational state |
Glossary, Critical Section |
MaxLogWriterTasks |
MAX_BACKUP_TASKS |
Maximum number of log writer tasks (determines the maximum number of log partitions) |
Changes only apply after a restart of the database. |
|
MaxMemoryAllocationSize |
MEMORY_ALLOCATION_LIMIT |
Maximum amount of working memory to be used by the database (in KB) |
Can be changed in ONLINE or ADMIN operational state |
- |
MaxParallelLiveCacheTraceFiles |
_MAX_MESSAGE_FILES |
Maximum number of trace files that the system can open simultaneously |
Changes only apply after a restart of the database. |
- |
MaxRetentionTime MinRetentionTime |
MAX_RETENTION_TIME MIN_RETENTION_TIME |
For SAP liveCache databases Specifies the time (in min) after which the garbage collectors start deleting the history data (undo log entries and links to them) The garbage collectors start deleting the history data in the following cases:
|
Can be changed in ONLINE or ADMIN operational state |
Concepts of the Database System, How Databases Log Data Changes |
MaxServerTasks |
MAXSERVERTASKS |
Maximum number of server tasks |
Changes only apply after a restart of the database. |
Concepts of the Database System, Tasks |
OfficialNodeName |
OFFICIAL_NODE |
Only for hot standby systems Virtual Server Name by which the cluster is addressed from the outside The system administrator initially assigns the Virtual Server Name to the computer on which the master database of the hot standby system is located. If the master database fails, then the name transfers to the standby database that takes on the master role. |
Changes only apply after a restart of the database. |
Glossary, Hot Standby |
OmsHeapThreshold |
OMS_HEAP_THRESHOLD |
For SAP liveCache databases Specifies how much working memory usage the OMS heap may use (in %) |
Changes only apply after a restart of the database |
- |
OmsSubHeaps |
OMS_HEAP_COUNT |
For SAP liveCache databases Specifies whether the OMS heap is divided into segments If the OMS heap is not divided into segments, the internal heap memory management locks the entire OMS heap whenever a database procedure requests or releases memory. |
Changes only apply after a restart of the database |
Concepts of the Database System, Working Memory Areas |
OmsVersionThreshold |
OMS_VERS_THRESHOLD |
For SAP liveCache databases Specifies how much working memory the OMS versions may use (in KB) |
Changes only apply after a restart of the database |
- |
RequestTimeout |
REQUEST_TIMEOUT |
Maximum time that a database session waits for a locked database object to be released (in s) If this time is exceed, the system sends a message to the waiting database session. The system then rolls back any changes that were previously executed within the respective transaction. |
Changes only apply after a restart of the database. |
Concepts of the Database System, Locks |
RequestUpdateStatisticsThreshold |
- |
If the system detects that the size of a table has changed by more than the value specified in this parameter (in percent) since the last statistics update, it marks the table as needing a statistics update. |
Can be changed in ONLINE or ADMIN operational state |
|
RowLockManagementStripes |
_ROW_RGNS |
Number of critical sections into which the row lock list is divided This division enables a parallel access to the lock list. |
Changes only apply after a restart of the database. |
Glossary, Critical Section |
RowLocksPerTransactionThreshold |
- |
Maximum percentage of all available locks that a transaction can hold before the system attempts a lock escalation. |
Can be changed in ONLINE or ADMIN operational state |
Glossary, Lock |
SequenceCacheSize |
SEQUENCE_CACHE |
Size of the sequence cache (in pages) |
Changes only apply after a restart of the database. |
Glossary, Sequence Cache |
SessionTimeout |
SESSION_TIMEOUT |
Timeout value for database sessions (in s) |
Can be changed in ONLINE or ADMIN operational state |
Glossary, Database Session |
SharedSQLCleanupThreshold |
SHAREDSQL_EXPECTEDSTATEMENTCOUNT |
Expected number of SQL statements to be stored in the shared SQL cache |
Can be changed in ONLINE or ADMIN operational state |
- |
SharedSQLCommandCacheSize |
SHAREDSQL_COMMANDCACHESIZE |
Maximum size of the Shared SQL cache (in KB) |
Can be changed in ONLINE or ADMIN operational state |
- |
TableLockManagementStripes |
_TAB_RGNS |
Number of critical sections into which the table lock list is divided This division enables a parallel access to the lock list. |
Changes only apply after a restart of the database. |
Glossary, Critical Section |
TraceBufferSize<task_type>Task |
TRACE_PAGES_<task_type> |
Each user kernel thread has its own trace area in the working memory. This parameter specifies the maximum number of pages that are available to the specified task type per user kernel thread. <task_type>:BUP (Backup Task, not yet implemented)|FS (Floating Service Task: either an event task or a Database Analyzer task)|GC (Garbage Collector)|LW (Log Writer)|PG (Pager)|SV (Server Task)|TI (Timer Task)|US (User Task)|UT (Utility Task) |
Changes only apply after a restart of the database. |
Concepts of the Database System, Tasks |
TransactionHistorySize |
TRANS_HISTORY_SIZE |
Specifies for how many long-running transactions the system stores information |
Changes only apply after a restart of the database. |
- |
TransactionHistoryThreshold |
TRANS_THRESHOLD_VALUE |
Time (in s) after which a transaction is considered long-running by the system |
Can be changed in ONLINE or ADMIN operational state |
- |
TransactionLockManagementStripes |
_TRANS_RGNS |
Number of critical sections into which the transaction lock list is divided This division enables a parallel access to the lock list. |
Changes only apply after a restart of the database. |
Glossary, Critical Section |
UpdateStatSampleAlgorithm |
UPDATESTAT_SAMPLE_ALGO |
Determines the algorithm that the database system uses to update the SQL optimizer statistics In SAP systems we recommend the value 1 for this parameter. |
Can be changed in ONLINE or ADMIN operational state |
SQL Optimizer, SQL Optimizer Statistics |
UseableCPUs |
- |
Restricts the number of available CPUs in a running system. The following constraint applies:1 ≤ UseableCPUs ≤ MaxCPUs To permanently configure the number of available CPUs, use the general database parameter MaxCPUs. |
Can be changed in ONLINE or ADMIN operational state |
|
UseBackupSecurityDescriptor |
USE_BACKUP_SECURITY_DESCRIPTOR |
Only on Microsoft Windows Specifies whether the system uses the Microsoft Windows Security Descriptor for backups to data carriers of the file type Caution If you want to carry out backups to a Samba share, set this parameter to NO. End of the caution. |
Changes only apply after a restart of the database. |
- |
UseDataCacheScanOptimization |
LRU_FOR_SCAN |
Specifies whether the whole data cache is used for scans |
Can be changed in ONLINE or ADMIN operational state |
- |
UseHashedResultset |
HASHED_RESULTSET |
Specifies whether the SQL optimizer uses temporary hash tables for creating aggregations. |
Can be changed in ONLINE or ADMIN operational state |
|
UseLobClustering |
CLUSTERED_LOBS |
Specifies whether the database system stores LOB values as clusters in the data area |
Can be changed in ONLINE or ADMIN operational state |
- |
UseSharedSQL |
SHAREDSQL |
Activates the Shared SQL function |
Parameter can be activated in ONLINE or ADMIN operational state but not deactivated |
Glossary, Shared SQL |
UseUnicodeColumnCompression |
- |
Specifies whether the system stores application data in columns of the type (VAR)CHAR UNICODE in newly created tables as UTF8 |
Can be changed in ONLINE or ADMIN operational state |
Glossary, Data Type |
VolumeFormattingMode |
FORMATTING_MODE |
The database system formats the volumes when the database is created. With this parameter, you specify the formatting type. Only use parallel formatting for data volumes of the FILE type that are located on different hard disks. |
Changes only apply after a restart of the database. |
Glossary, Volume |