[Srv]
|
Description
|
Factory Value
|
Access Mode
|
---|---|---|---|
AbortTimeOut
|
Specifies the time in minutes after an idle transaction is aborted; negative or zero value means infinite.
|
120
Unit: 1 min
|
RW/
Startup |
AdaptiveRowsPerMessage
|
This parameter takes the average number of rows returned to the client as the rows per message value. The start value grows as more rows are fetched. If set to no, the RowsPerMessage parameter value is used. That is also the default value.
|
yes
|
RW/
Startup |
AllowConnect
|
If set to no, only connections from solidDB® Remote Control (solcon) or solidDB® SQL Editor (solsql) are allowed
|
yes
|
RW/
Startup |
At
|
This parameter can be used specify commands for automating administrative tasks, such as executing operating system commands, creating backups, checkpoints, and database status reports.
The syntax of the value for this parameter is the following:
At = At_string At_string ::= timed_command [ ,timed_command ]
timed_command ::= [ day ] HH:MM argument
day ::= sun | mon | tue | wed | thu | fri | sat
For example:
At = 20:30 makecp, 21:00 backup, sun 23:00 shutdown
If you specify a backup, the default backup directory is the one set with the General.BackupDirectory parameter.
If the day is not given, the command is executed daily.
There is no factory value for this parameter.
For more information about entering time commands, including a list of the available commands and their arguments, see Entering timed commands.
|
(no factory value)
|
RW
|
ConnectionCheckInterval
|
This parameter specifies the number of seconds between connection status checks in thread/client mode.
To use this parameter, you must set the Srv.ReadThreadMode parameter to 0 and the Srv.Threads to a large enough value to accommodate the threads in your environment.
When the Srv.ReadThreadMode parameter is set to 2 (default), the server does not detect a broken connection until it tries to write something back to the client.
|
10
Unit: seconds
|
RW/
Startup |
ConnectTimeOut
|
Specifies the continuous idle time in minutes after a connection is dropped; negative or zero value means infinite.
Note The value set with this parameter is not effective for the SMA handshake connection that is used to pass the shared-memory segment handle to the SMA driver.
|
480
Unit: 1 min
|
RW/
Startup |
DatabaseSizeReportInterval
|
When the database size exceeds the limit defined with this parameter, the system generates a report file. This parameter gives the delta after which the next report is printed. The minimum delta value is 1 MB. The report file name is repdb<mb>MB.dbg.
This parameter is useful, for example when tracing unexpected database size growth.
If you leave this parameter to its default value 0, no reports are generated. The minimum non-zero value for this parameter is 1 MB.
|
0 MB
|
RW/
Startup |
DisableOutput
|
Disables generation of the solmsg.out and the solerror.out files. For details on these files, see Viewing error messages and log files. To disable file generation, this parameter must be included in the solid.ini file and set to yes. If this parameter is set to no or it is not included in the solid.ini file, the log files are generated.
|
no
|
RO
|
Echo
|
If set to yes, contents of solmsg.out file are displayed also at the server's command window.
|
no
|
RW/
Startup |
ExecRowsPerMessage
|
This parameter specifies how many result rows are sent (prefetched) to the client driver in response to the SQLExecute call with a SELECT statement. The result rows are subsequently returned to the application with the first SQLFetch calls issued by the application. The default value of 2 allows for prefetching of single-row results. If your SELECT statements usually return larger number of rows, setting this to an appropriate value can improve performance significantly.
See also the RowsPerMessage parameter.
|
2
|
RW/
Startup |
ForceThreadsToSystemScope
|
This parameter applies only to symmetric multiprocess (SMP) Solaris operating systems, in which the default scope provided by the threads of the runtime library can be set to process scope, system scope, or light weight process (lwp) scope. In Solaris environments, threads are lightweight processes.
The value yes can improve the server's performance in a multi-CPU machine significantly. The actual performance improvement depends on how evenly the workload is already spread across your CPUs. If this parameter is set to no, you should get slightly better performance in single-CPU systems.
When this parameter is set to yes, it forces lwp threads to be run in system scope, instead of process scope. It also allows Solaris to schedule solidDB® threads on any available CPU. This reduces bottlenecks and enhances the parallelization of operations, including I/O. For more information on lwp, see Solaris operating system documentation.
|
Solaris: yes
Other environments: no
|
RW/
Startup |
HealthCheckEnabled
|
When the parameter is set to yes, a periodical check is performed to detect a stalled server due to, for example, unexpected operating system stalls or software errors.
The check uses a timeout-based server deadlock detection algorithm that checks certain critical low-level concurrent programming synchronization objects (mutexes).
If a deadlock is detected, the server process terminates with an error and a message is printed to solerror.out.
For example in High Availability (HotStandby) configurations, a failover can be enforced upon the detection of a server deadlock.
Note This parameter is not related to transaction-level deadlock detection mechanisms.
|
no
|
RW/
Startup |
HealthCheckInterval
|
This parameter sets the interval of the server deadlock check.
Unit: seconds
|
60
|
RW
|
HealthCheckTimeout
|
This parameter sets the deadlock detection timeout time.
The factory value is high enough to escape false errors. If faster detection is needed, set the parameter to a lower value.
Unit: seconds.
|
60
|
RW
|
InifileLineSplitting
|
Defines whether in solid.ini configuration file lines that are longer than 79 characters are split into multiple lines when server saves the file.
For example, if you create comments longer than 79 characters, the server splits the comments on separate lines using a backslash (\) at the end of the line but without adding a comment marker (;) on the new line. The server handles the lines that have been split in this way; however, applications such as watchdogs might see the file as corrupted and thus fail.
Value no means lines are never split.
|
yes
|
RW/
Startup |
KeepAllOutFiles
|
If this parameter is set to yes, the solidDB® message log (solmsg.out) and trace files are not overwritten with new contents. Instead, when a file limit is reached, a new file is created with an incremented file name number postfix. The starting value of the postfix is set by using parameters Srv.TraceBackupFileNum and Srv.SolmsgBackupFileNum.
|
no
|
RO
|
LocalStartTasks
|
Number of server's internal tasks that execute the local background statements that were started with command START AFTER COMMIT (without FOR EACH REPLICA).
Valid values range from 1 - 100.
Note In this context task refers to solidDB®'s internal tasks, not thread or task as used in some Real-Time Operating Systems. A task is an operation that has to be executed, such as checkpoint, backup, or SQL statement.
In this case, you can have 1 to N tasks that execute the background operations. More tasks mean that background tasks reserve more resources and are handled faster, and that other operations (for example, interactive ones) will get fewer resources and be handled more slowly.
|
2
|
RW/
Startup |
MaxBigTaskInterval
|
This parameter (MAXimum BackGround TASK MaxBgTaskInterval INTERVAL) tells the server the maximum length of time to wait before checking whether internal administrative tasks that are “sleeping” should be “awakened”. The units are seconds.
For example, if a connection has been broken or disconnected, this parameter specifies the maximum length of time that the server will wait before noticing that the connection is gone. This time is IN ADDITION TO whatever time is required for the underlying communication layer to detect that the connection is broken. For example, if you have a Connect Timeout of 100 seconds and a MaxBgTaskInterval of 50 seconds, then you may have to wait up to 150 seconds before a broken connection is detected and no longer counted as one of the connections.
You may want to set or adjust this parameter if you get errors similar to the following:
Error 08004: [Solid] [SOLID ODBC Driver]
[SOLID]SOLID Server Error 14507: Maximum number of licensed user connections exceeded This parameter only applies to the server's own internal administrative tasks. It does not affect the scheduling of user tasks.
Note MaxBgTaskInterval applies to all server administration tasks, regardless of each task’s priority. Even when a high priority task is running, the server will check the low-priority tasks at the specified intervals.
Setting MaxBgTaskInterval to a small enough value may reduce performance and may reallocate some time from high-priority tasks to low-priority tasks. This can happen in systems where low-priority connections are not checked often enough to notice that they have been disconnected. However, because the parameter only affects server administrative tasks, not user tasks, the effect is generally small.
|
2 (seconds)
|
RW/
Startup |
MaxConstraintLength
|
This parameter controls the maximum number of bytes that the server will search through in a string, for example in WHERE clauses such as:
WHERE LOCATE(sought_string, column1) > 0;
For example, if the value is 1024, ASCII character strings are searchable up to 1024 characters and Unicode character strings are searchable up to 512 characters (1024 bytes).
This parameter applies to strings that have the following data types:
CHAR(#) VARCHAR(#)
It does not apply to strings that have the data type(s):
LONG VARCHAR
The minimum valid value is 254. If you specify a smaller number, the server will still search the first 254 bytes. Although you can use any value from 254 to 2G-1, practical values are generally in the range of a few kilobytes, like 1024, or 8192.
|
254 (254 bytes = 254 ASCII characters, or 127 Unicode characters)
|
RW
|
MaxOpenCursors
|
The maximum number of cursors that a database client can have simultaneously open.
|
1000
|
RW/
Startup |
MaxRPCDataLen
|
This allows you to specify the maximum string length of a single SQL statement sent to the server. This is particularly useful if you are sending CREATE PROCEDURE commands that are longer than 64K. The value should be between 64K (65536) and 1024K (1048576). If the value is less than 64K, the server will use a minimum of 64K.
|
512K (524288)
|
RW/
Startup |
MaxStartStatemenst
|
Maximum number of simultaneous "uncommitted" START AFTER COMMIT statements. Valid values range from 0 - 1000000.
|
10000
|
RW/
Startup |
MaxUsers
|
Defines the maximum number of connections to solidDB®.
When the number of maximum users has been exceeded, error 14507 is issued and you can connect to solidDB® only with solidDB® Remote Control (solcon).
Value 0 means that the maximum number of connections is not restricted
|
0
|
RW/
Startup |
MemoryReportDelta
|
This parameter defines how much memory allocations must increase or decrease compared to the previous message before the new message is printed to solmsg.out.
|
20 MB
|
RW/
Startup |
MemoryReportLimit
|
This parameter defines the minimum size for memory allocations after which reporting to solmsg.out is done.
|
0 (no reporting)
|
RW/
Startup |
MemorySizeReportInterval
|
When the memory size exceeds the limit defined with this parameter, the system generates a report file. This parameter defines the delta after which the next report is printed. The minimum delta value is 1 MB. The report file name is repmem<mb>MB.dbg.
This is parameter is useful, for example when tracing unexpected memory growth in the server.
If you leave this parameter to its default value 0, no reports are generated. The minimum non-zero value for this parameter is 1 MB.
|
0 MB
|
RW/
Startup |
MessageLogSize
|
The maximum size of the solmsg.out file in bytes.
|
1 MB
Unit: 1byte k=KB m=MB
|
RW/
Startup |
Name
|
Specifies the informal name of the server, equivalent to the -n command line option.
|
|
RW/
Startup |
NetBackupRootDir
|
Sets the root directory for the network backups in NetBackup Server. The path is relative to the working directory.
|
The working directory
|
RW
|
ODBCDefaultCharBinding
|
Defines the binding method for character data types. The options are:
▪raw - no data conversion takes place between solidDB® server and the client
The value raw can be used when you want your database to use the binding used in the6.3 or earlier versions of solidDB®.
▪locale - the current client locale setting is used, also if set by the client system locale: - the current client setting are overridden with a default locale set of the client system The driver calls setlocale() with an empty string which effectively searches for the locale setting set in the system.
For example, in Linux environments, the environmental variable LC_CTYPE is checked first and if that is not defined, the environmental variable LANG is searched.
▪locale:<locale name> - the current client systems setting are overridden and the given locale is used
The convention for <locale name> depends on the operating system.
For example, in Linux environments, the locale name for the code page GB18030 in Chinese/China is zh_CN.gb18030. In Windows environments, the locale name for Latin1 code page in Finnish/Finland is fin_fin.1252.
▪UTF8 - UTF-8 binding is enforced regardless of the locale set in the client-side system
The factory value depends on the value of the parameter General.InternalCharEncoding:
▪If General.InternalCharEncoding is 'raw', ODBCDefaultCharBinding is also 'raw'.
▪If General.InternalCharEncoding is 'UTF8', ODBCDefaultCharBinding is 'locale:'.
|
raw
See description
|
RW/
Startup |
PessimisticTableUseNFetch
|
Pessimistic table locks are used to prevent other sessions from adding, editing, or deleting any records or placing any record or table locks on a given table. Table locks block other record or table lock attempts, but do not block any reads of the locked table.
If pessimistic tables are used, they force the RowsPerMessage value to 1 if the query locks any rows. You can enable the RowsPerMessage for pessimistic tables by enabling the PessimisticTableUseNFetch parameter. By default, it is disabled.
|
no
|
RW/
Startup |
PrintMsgCode
|
Causes a unique 8-character message code to be inserted before each status and error message in the message log files (solmsg.out and solerr.out).
|
no
|
RW/
Startup |
ProcessMemoryCheckInterval
|
The process size limits are checked periodically. The check interval is set with the ProcessMemoryCheckInterval parameter. The interval is given in milliseconds.
The minimum non-zero value is 1000 (ms). Only values 0 or 1000 or above 1000 (1 second) are allowed. If the given value is above 0 but below 1000, an error message is given.
The factory value is 0, that is, process size checking is disabled.
The ProcessMemoryLimit and ProcessMemoryCheckInterval parameters are interlinked; if the ProcessMemoryCheckInterval parameter is set to 0, the ProcessMemoryLimit parameter is not effective, that is, there is no process memory limit.
See also parameters ProcessMemoryLowPercentage and ProcessMemoryWarningPercentage.
|
0
|
RW
|
ProcessMemoryHysteresisPercentage
|
As the amount on memory used crosses different boundaries specified with, for example, the ImdbMemoryLowPercentage or the ProcessMemoryLimit parameter, system events are given. The event behavior expresses hysteresis in a way that the value triggering the BELOW event is somewhat lower than the specified value triggering the ABOVE event. The difference can be, for instance, 5%. As a result, the number of system events is not too large if the amount of memory alternates rapidly just above and below the specified boundaries. The ProcessMemoryHysteresisPercentage parameter is used to set the difference as a percentage value.
|
5
|
RW
|
ProcessMemoryLimit
|
This parameter specifies the maximum amount of virtual memory that can be allocated to the in-memory database process.
When this limit is exceeded, the server gives an error message and accepts admin commands only. The limit can be changed dynamically.
The ProcessMemoryLimit and ProcessMemoryCheckInterval parameters are interlinked; if the ProcessMemoryCheckInterval parameter is set to 0, the ProcessMemoryLimit parameter is not effective, that is, there is no process memory limit.
Note You should not set the Srv.ProcessMemoryLimit and Srv.ProcessMemoryCheckInterval parameters when using SMA. If you need to limit the memory the SMA uses, use the SharedMemoryAccess.MaxSharedMemorySize parameter.
|
1G
Unit: 1 byte, G=GB, M=MB, K=KB
|
RW
|
ProcessMemoryLowPercentage
|
The ProcessMemoryLowPercentage parameter sets a warning limit for the total process size. The limit is expressed as percentage of the ProcessMemoryLimit parameter value.
Prior to exceeding this limit, you have exceeded the warning limit defined by using the ProcessMemoryWarningPercentage parameter and received a warning. When the ProcessMemoryLowPercentage limit is exceeded, a system event is given.
The ProcessMemoryLowPercentage parameter value is automatically checked for consistency. It must be higher than the ProcessMemoryWarningPercentage parameter value.
See also parameters ProcessMemoryLimit, ProcessMemoryCheckInterval , and ProcessMemoryWarningPercentage.
|
90
|
RW
|
ProcessMemoryWarningPercentage
|
The ProcessMemoryWarningPercentage parameter sets the first warning limit for the total process size. The warning limit is expressed as percentage of the ProcessMemoryLimit parameter value. When the ProcessMemoryWarningPercentage limit is exceeded, a system event is given.
The ProcessMemoryWarningPercentage parameter value is automatically checked for consistency. It must be lower than the ProcessMemoryLowPercentage parameter value.
See also parameters ProcessMemoryLimit, ProcessMemoryCheckInterval , and ProcessMemoryLowPercentage.
|
80
|
RW
|
ReadThreadMode
|
This parameter controls the number of threads that the server uses to service client requests. If the value is 0, the server uses the number of threads specified with the parameter Threads. If the value is 2, the server creates a separate thread for each client. Using more threads will generally improve performance, but also requires more memory.
This parameter only controls the number of threads serving client requests. It does not affect the number of threads doing other work within the server.
Some operating systems may limit the maximum number of threads allowed, and setting this parameter's value to 2 may cause the server to request more threads than the OS allows. If you try to exceed the number of threads allowed, you will get the following type of error:30146 Failed to create thread 'dnet_clientthread'
|
2
|
RW/
Startup |
RemoteStartTasks
|
Number of Replica server's internal tasks inside the server that execute the remote background statements started at Master with command START AFTER COMMIT... FOR EACH REPLICA. Valid values range from 1 - 100.
Note In this context task refers to the internal tasks of the solidDB® server, not thread or task as used in some Real-Time Operating Systems. A task is an operation that has to be executed, such as checkpoint, backup, or SQL statement.
|
2
|
RW/
Startup |
ReportInterval
|
Enables automated generation of reports at the given interval (in seconds).
Automated reports are named reptimestamp>.dbg and output to the solidDB® working directory.
Value 0 means that reports are not generated automatically.
|
0
|
RW
|
RowsPerMessage
|
Specifies the number of rows returned from the server in one network message when an SQLFetch call is executed (and there are no prefetched rows).
See also the ExecRowsPerMessage configuration parameter.
|
100
|
RW/
Startup |
Silent
|
If set to yes, no output is generated to the server’s command window. Only license information is displayed.
|
no
|
RW/
Startup |
SolmsgBackupFileNum
|
This parameter defines the starting digit of the message log file (solmsg.out) name postfix, if the KeepAllOutFiles parameter is set to yes.
For example, if the value is set to 5, the solmsg.out files are named as follows:
solmsg5.out
solmsg6.out solmsg7.out ... Valid values range from 0 to 999999.
|
0
|
RW/
Startup |
StackTraceEnabled
|
The StackTraceEnabled parameter controls the stack trace functionality upon an assertion failure or a signal caused by server malfunction. When set to yes, the stack trace information is output to ssstacktrace-<process_id>-<thread_id>.out file.
The following signals invoke the stack traces output automatically:
▪SIGSEGV
▪SIGILL
▪SIGBUS
▪SIGTRAP
▪SIGSYS
▪SIGEMT
The stack traces information is produced only about the thread that received the signal.
Additionally, you can generate the stack traces information for all currently running threads by sending the server the SIGUSR1 signal.
Note The stack trace feature is not available on Windows systems.
|
no (Linux 64-bit)
yes (Linux 32-bit and UNIX systems)
|
RW/
Startup |
StandardDateTimeFormat
|
By default, solidDB® uses the ISO/IEC/ANSI standard date representation, which is also the standard date literal format in SQL. The date is represented as shown in the timestamp example below:
2008-10-15 09:29:40
When set to no, the message log files (solmsg.out) use a date format such as 15.10 09:29:40. The solerror.out file uses a date format such as Mon Oct 22 15:16:35 2007.
|
yes
|
RO
|
StatementMemoryTraceLimit
|
This parameter switches on tracing for statements that have allocated memory over the defined value. These statements are put into the peak memory usage list. The peak memory list is printed to report file. Statements that use memory over the defined limit are also printed to the solmsg.out file.
|
0 MB
|
RW/
Startup |
Threads
|
If the Srv.ReadThreadMode parameter is set to 0, this parameter specifies the number of concurrent threads that the server uses to process user requests. The helper threads, such as I/O threads, are not included in the count.
If the value of Srv.ReadThreadMode is other than 0, the value of this parameter is insignificant, as the server controls the number of threads automatically.
|
5
|
RW/
Startup |
TraceBackupFileNum
|
The starting value of the trace file name postfix appended to the file name if the KeepAllOutFiles parameter is set to yes.
Valid values range from 0 to 999999.
|
0
|
RW/
Startup |
TraceLogSize
|
This parameter allows you to limit the maximum size of the trace log file. The size is specified in bytes; for example, TraceLogSize=10000 limits the size of the trace log file to 10000 bytes. The trace log file is the file to which the server writes information when you turn on monitoring.
For information about turning on monitoring, see the description of ADMIN COMMAND 'monitor...' in ADMIN COMMAND and the -m command-line option in solidDB® command-line options.
Monitoring uses the file named soltrace.out for output. After reaching the maximum size, the following takes place:
▪solidDB® deletes any existing file named soltrace.bak;
▪solidDB® renames the current soltrace.out file to soltrace.bak; and
▪solidDB® starts a new soltrace.out file.
|
100 m
Unit: 1 byte k=KB m=MB
|
RW/
Startup |
TraceSecDecimals
|
Number of second decimals in trace outputs. Allowed values are from 0 to 3.
|
3
|
RW/
Startup |