Entering content frame

Function documentation Support Database Parameters Locate the document in the library structure

Use

Most support database parameters are for troubleshooting. In normal operation, it is not necessary to change the support database parameters. Changing support database parameters requires very detailed knowledge of the database system. For information about other database parameters, see General Database Parameters, Special Database Parameters and liveCache Database Parameters.

Caution

Changes to database parameters do not take effect until the database instance is restarted.

With the Database Manager CLI, you can change some database parameters even while the system is running.

Support Database Parameters

Parameter

Meaning

How can I change this Parameter?

_BACKUP_HISTFILE

Name of the history file for data and log backups

Parameter cannot be changed

_BACKUP_MED_DEF

Name of the file containing the definitions of the backup media

Parameter cannot be changed

_EVENTFILE

Name of the file in which the database system logs events

Restart required after change

_EVENTFILE

Size of the file in which the database system logs events

The database system overwrites this file cyclically.

Restart required after change

_IDXFILE_LIST_SIZE

Number of temporary intermediate result files in the case of parallel indexing

The database system indexes large tables using multiple server tasks. These server tasks write their results in temporary files. When the number of these files reaches the value of this parameter, the database system has to merge the files before it can generate the actual index. This results in a decline in performance.

Restart required after change

_IOPROCS_PER_DEV

Number of threads that the database system can use for asynchronous I/O operations (per data volume and system)

Restart required after change

_KERNELDIAGFILE

Name of the log file of the kernel

Restart required after change

_KERNELDUMPFILE

Name of the dump file that is written by the kernel when the system crashes

To this file, the database system writes, among other things, the contents of the data cache and the converter. Size your system so that sufficient memory space is available for this file (approximately CACHE_SIZE + 10%). Note also parameter DIAG_HIST_NUM.

Restart required after change

_KERNELTRACEFILE

File in which the kernel writes the trace entries

The kernel only writes trace entries if you previously activated the database trace.

Restart required after change

_MAXEVENTTASKS

Maximum number of event tasks

The database system uses event tasks both for Event Dispatcher and for the DBM commands event_wait, event_receive, event_available, auto_extend and auto_update_statistics. These commands can be active simultaneously for this database instance.

The default value of the Database Manager for this parameter is 2.

Restart required after change

_MAXTASK_STACK

Size of the stack that is used by the user tasks

Restart required after change

_MINREPLY_SIZE

Minimum size of the memory that is available in a package (shared memory segment) for the reply

0: For the application as well as for the kernel, the whole package is available for the request/reply.

Restart required after change

_MULT_IO_BLOCK_CNT

Block size used for writing data pages from the data cache to the data area and for writing log pages from the log queue to the log area

The optimum block size depends on the hardware and software that is used.

Restart required after change

_PACKET_SIZE

Size of the packages (shared memory segments) into which SQL statements and data are transmitted

A package consists of a request part and a reply part.

Restart required after change

_READAHEAD_BLOBS

Number of pages from which large LONG values are imported in advance by additional server tasks

If a LONG value is too large to be transferred to the client in a single request package, the database system splits it into several request packages. To increase performance, server tasks can import further parts of the LONG value while the first request package is being sent.

Parameter can be changed with the Database Manager CLI while in operation

_RESTART_TIME

Minimum time between two savepoints (in seconds); corresponds to the time that is required for a restart after a system crash

Beyond this, the database system always writes a savepoint in the following cases:

·        When you have created an index

·        When the number of pages released by a savepoint is larger than the number of free blocks in the data volumes

Restart required after change

_RTEDUMPFILE

File in which the kernel writes information about the runtime environment if the system crashes

Restart required after change

_SERVERDB_FOR_SAP

Specifies whether the database instance is being used in an SAP system

Parameter can be changed with the Database Manager CLI while in operation

_TASKCLUSTER_01 to _03

These parameters describe how the database system distributes user tasks to the threads. Only change these parameters after consulting support.

Restart required after change

_USE_ASYNC_IO

Specifies whether operating system functions or special I/O threads are used for asynchronous I/O operations

Restart required after change

_USE_IOPROCS_ONLY

YES: I/O operations are executed only by special I/O threads

NO: the kernel decides whether a task is to execute an I/O operation itself or give the I/O operation to a special I/O thread

Parameter can be changed with the Database Manager CLI while in operation

ALLOW_MULTIPLE_SERVERTASK_UKTS

Specifies whether the database system distributes the server tasks to the available user kernel threads or if they all run in the same user kernel thread

Restart required after change

ENABLE_SYSTEM_TRIGGERS

Defines whether system triggers are called when the database is restarted

Restart required after change

EXPAND_COM_TRACE

Specifies whether memory space is reserved when COM trace files are created

Restart required after change

INIT_ALLOCATORSIZE

Initial size of the main memory that is reserved at the start of a database session

Restart required after change

LOAD_BALANCING_DIF

Only for load balancing; specifies by how much longer the task to be moved had to wait in its user kernel thread than the task that has been waiting the longest in the target user kernel thread (as a percentage)

Restart required after change

LOAD_BALANCING_EQ

For load balancing; specifies which time delay should be regarded as equal when comparing waiting tasks (as a percentage)

Restart required after change

MAX_SERVERTASK_STACK

Maximum size of the stack that is used by the server tasks

Restart required after change

MAX_SPECIALTASK_STACK

Maximum size of the stack that is used by special tasks (all except user tasks and server tasks)

Restart required after change

MAXPAGER

Maximum number of pagers; calculated by the database system from MAXDATAVOLUMES, among other parameters

Not changeable

You can override this parameter with XP_MAXPAGER.

MAXVOLUMES

Maximum number of data and log volumes including mirrored volumes; calculated by the system

Restart required after change

MINI_DUMP

Defines whether the system writes the memory dumps knlmini.dmp and srvmini.dmp (for postmortem debugging) and how much information these dumps contain

DISABLED: the system does not write any memory dumps

NORMAL (default value): the system writes all stacks and system handle information in the memory dump

FULL: the system writes all stacks and data segments in the memory dump

Restart required after change

OMS_STREAM_TIMEOUT

Maximum wait time for all database sessions that may elapse until the reply to an OMS stream request to a client (in seconds)

Parameter can be changed with the Database Manager CLI while in operation

OPTIM_CACHE

Defines whether the database system determines the search strategy only once or every time it executes a parsed SQL statement

For prepared statements with parameters, it may be sufficient to determine the search strategy only once.

Parameter can be changed with the Database Manager CLI while in operation

OPTIMIZE_OPERATOR_JOIN

Defines whether the database system uses optimized join implementation (saves resources)

Parameter can be changed with the Database Manager CLI while in operation

SET_VOLUME_LOCK

Specifies whether the database system sets a lock when a volume is added, which prevents the same volume from being added again later

In the following cases it may make sense to set this parameter to NO:

·        NFS-mounted volume

·        Hot standby system

Restart required after change

SHOW_MAX_STACK_USE

Logs the maximum stack usage of each task in the kernel log

Only set this parameter to YES for diagnosis purposes as it compromises performance.

Restart required after change

SUPPRESS_CORE

Defines whether the database system suppresses core dumps of the kernel

Restart required after change

SYMBOL_DEMANGLING

Defines whether the database system performs C/C++ demangling

Restart required after change

UPDATESTAT_PARALLEL_SERVERS

Defines how many parallel server tasks the database system uses for updating the SQL Optimizer statistics

Value -1: update of statistics is done sequentially
Value 0: database system decides how many parallel server tasks should be used

Value n>0: database system uses a maximum of n parallel server tasks

Parameter can be changed with the Database Manager CLI while in operation

USE_SYSTEM_PAGE_CACHE

Defines whether the database instance uses the system page cache for buffering memory pages that are no longer required

Restart required after change

 

 

Leaving content frame