The SCREEN command displays a screen of PDE configuration information from the PDE Control GDO. Some of this information can be modified using the variable = setting and WRITE commands.
- Each screen displays groups of related control fields. To change the values of modifiable fields, use the variable = setting command, identifying the field either by the full field name or by the alphanumeric identifier that appears next to the field in the screen output. Fields lacking alphanumeric identifiers should not be modified.
- Scripts that change field values should use the full variable names, rather than the alphanumeric identifiers.
- Entering the screen command alone, without a screen name, redisplays the current screen (the one that was most recently displayed).
The ctl screens are described in the sections that follow:
The DBS screen displays parameters that control how the database software responds to unusual conditions.
Example: SCREEN DBS output
>screen dbs (0) Minimum Node Action: Clique-Down (1) Minimum Nodes Per Clique: 1 (2) FSG cache Percent: 90 (3) Clique Failure: Clique-Down (4) Cylinder Read: On (5) Restart After DAPowerFail: On (6) Cylinder Read Ageing Threshold: 0 (7) Maximum Fatal AMPs: 0 (7) TIM FSG cache Percent 50
The DBS screen contains the following control fields.
|Minimum Node Action||Determines what action to perform when a clique contains fewer than the Minimum Nodes Per Clique field.
|Minimum Nodes Per Clique||Specifies the number of nodes required for a clique to operate. (Inactive hot standby nodes are not considered.) If a clique contains fewer than this number of nodes when Teradata Database is started, the Minimum Node Action setting determines what action to take.
Minimum Nodes Per Clique does not affect cliques in which all nodes are running properly, including “single-node cliques” such as AMP-less Channel nodes. These cliques are exempt from the Minimum Node Action, regardless of the Minimum Nodes Per Clique setting.
Changes to this value do not take effect until Teradata Database is restarted.
When Teradata Database starts, the system determines the minimum number of nodes required for each clique, based on the number of vprocs, nodes, and available memory in the clique. For systems containing cliques of different sizes, Teradata determines a minimum node requirement for each clique, then considers the largest value as the minimum node requirement for all cliques in the system.
Choosing a value for Minimum Nodes Per Clique involves a trade off between performance and system availability. When one or more nodes in a clique fail, the AMPs assigned to the failed nodes migrate to the remaining nodes in the clique. System performance can degrade when some nodes handle more vprocs than other nodes.
Setting a Minimum Nodes Per Clique value allows you to define at what point it is more efficient for the system to consider a partially disabled clique to be entirely unavailable, allowing the Teradata fallback logic to compensate for the problem. Note that running a system with fallback limits some functions, which should be a consideration when choosing an appropriate value for this setting.
|FSG Cache Percent||Specifies the percentage of available memory that can be used for the database file segment cache. The available memory is the memory that remains beyond the memory required to run utilities and the Teradata Database programs.
Changes to this value do not take effect until Teradata Database is restarted.
Setting the value to zero resets FSG Cache Percent to the factory default. You cannot disable this setting.
|TIM FSG Cache Percent||Specifies the size of the hot cylinder cache as a percentage of FSG cache.
This field is visible only on platforms that support the Teradata Intelligent Memory feature.
|Clique Failure||Determines what to do when a clique is down.
|Cylinder Read||Allows full-table scan operations to run more efficiently by reading the majority of data blocks on a cylinder with a minimal number of I/O operations.
A sector is the smallest unit of addressable disk storage. In the Teradata Database file system, a fixed number of sectors are grouped together to form a logical cylinder. Each cylinder contains several data blocks. A data block stores rows from a single database table, and spans several contiguous sectors within the cylinder. A single data block is the smallest unit of I/O for the Teradata Database file system. Every data block that is read incurs a disk I/O operation.
The majority of data blocks on a single cylinder usually store rows from a single table. Cylinder Read allows the file system to read multiple data blocks from a cylinder with a single I/O. For database operations that require scanning a complete table, this is a more efficient way to read data than performing a separate I/O for every data block of the table.
|Restart After DAPowerFail||Allows you to select whether or not to restart the Teradata Database after a disk array AC power failure.
|Cylinder Read Ageing Threshold||Specifies the maximum amount of FSG cache (in MB) that is used for segments loaded by Cylinder Read. A value of 0 represents the factory default, which is 96 MB for systems that use large cylinders or 16 MB for systems that use small cylinders.|
|Maximum Fatal AMPs||Specifies the maximum number of FATAL AMPs at or above which Teradata Database will not restart.
The default value is 0, which allows the database to restart even with FATAL AMPs. This is intended mainly for systems that have fallback protection. Even for systems without fallback protection, some queries can be completed when AMPs are offline if the queries do not access rows that are on the FATAL AMP.
For sites with critical tables that do not use fallback protection, this can be set to 1 to prevent the database from restarting until the problem that caused the FATAL AMP is resolved. For additional protection against running with FATAL AMPs, when this setting is 1, set Minimum Node Action and Clique Failure to DBS-DOWN.
The Debug screen is used for internal debugging of PDE and the Teradata Database.
Example: SCREEN DEBUG output
> scr debug (0) Start DBS: On (1) Break Stop: Off (2) Start With Logons: All (3) Start With Debug: Off (4) Save Dumps: Off (5) Snapshot Crash: Off (6) Maximum Dumps: 5 (7) Start PrgTraces: Off (8) Restart Dump Type: Selective (9) UDF Debugging: Off
The Debug screen contains the following control fields.
|Start DBS||Determines whether Teradata Database is started automatically when PDE starts.
|Break Stop||Controls whether the Teradata Database restarts automatically or stops for the system debugger when a fatal error occurs.
|Start With Logons||Controls whether logons are enabled or not.
The following values apply:
Changes to this setting take effect after the next Teradata Database restart. Start With Logons commands issued from the Supervisor window of Database Window have the same effect, but do not require a database restart. For more information, see Database Window (xdbw).
|Start with Debug||Halts the database software startup until after the system debugger has been attached.
|Save Dumps||Specifies whether database dumps are to be loaded into the database.
The following values apply:
|Snapshot Crash||Specifies whether Teradata Database continues to run after a snapshot dump.
|Maximum Dumps||Applies to a per-node basis and is meaningful only for database dumps. Controls the maximum number of crash dumps that will be saved. A value of -1 means the system will save as many dumps as can fit on the disk containing the dump directory.
The default is 5. Setting this field to 0 disables database dumps.
|Start PrgTraces||Allows you to save or not save prgtraces to files.
|Restart Dump Type||Specifies the type of memory dump to be performed for fatal database errors. There are two types:
Selective dumps are smaller, and can be collected more quickly. This provides faster Teradata Database restarts.
|UDF Debugging||Permits debugging sessions in Teradata Database. You must use a debugger supplied by Teradata to debug UDFs and external stored procedures running in Teradata Database.
Teradata offers command-line and Eclipse (Studio) plug-in debuggers for C/C++ and Java. For more information on these debuggers, see “C/C++ Command-line Debugging for UDFs” in SQL External Routine Programming, Teradata Debugger for C/C++ UDF, and Teradata Debugger for Java UDF.
The default is Off.
Teradata Database resource usage (ResUsage) statistics are collected by the Resource Sampling Subsystem (RSS). These statistics are logged to special tables in the database.
The RSS screen allows you to specify the rates of ResUsage data logging.
Example: SCREEN RSS output
>screen RSS (0) Node Logging Rate: 600 sec RSS Table Logging Enable (1) SPMA : On (2) IPMA : Off (3) SCPU: Off (4) SVPR : On (5) IVPR : Off (6) SLDV: On (7) SHST: Off (8) SPDSK: On (9) SVDSK: On (A) SAWT: On (B) SPS : On RSS Summary Mode Enable Summarize SPMA: Off Summarize IPMA : Off (C) Summarize SCPU : Off (D) Summarize SVPR: Off (E) Summarize IVPR : Off (F) Summarize SLDV : Off (G) Summarize SHST: Off (H) Summarize SPDSK: Off (I) Summarize SVDSK: Off (J) Summarize SAWT: Off Summarize SPS : Off
The RSS screen contains the following control fields.
|Node Logging Rate||The interval (in seconds) between writing statistics to the database.
The default is 600.
|RSS Table Logging Enable||Controls whether logging is enabled to the various ResUsage tables.
Only the SPMA table has logging enabled by default.
|RSS Summary Mode Enable||When logging is enabled for certain ResUsage tables, multiple rows of resource usage data are written during each logging period. Summary mode reduces the amount of data collected per logging period by causing the RSS to store a single summary row per type per node, instead of one row per logging entity.
For example, if regular logging is enabled for the SCPU table, separate rows storing statistics for every CPU are written during each logging period. If summary mode is enabled, only a single row is written for each node, regardless of the number of CPUs in that node. The single row includes summary data for all node CPUs.
Similarly, if regular logging is enabled for the SVPR table, separate rows are written for every individual vproc. If summary mode is enabled for this table, one row is written for each vproc type (AMP, PE, and others).
RSS Summary Mode is effective for a table only if RSS Table Logging is also enabled.
- To run ResUsage reports, you must select the logging rate and enable the appropriate tables for logging.
- Fields without alphanumeric identifiers should not be modified.
- RSS aligns the logging periods to the clock on the top of every hour. Therefore, these values must divide evenly into 3600 seconds (one hour). The following set of values are legal for RSS logging rates: 0, 1, 2, 3, 4, 5, 6, 8, 9, 10, 12, 15, 16, 18, 20, 24, 25, 30, 36, 40, 45, 48, 50, 60, 72, 75, 80, 90, 100, 120, 144, 150, 180, 200, 225, 240, 300, 360, 400, 450, 600, 720, 900, 1200, 1800, 3600.
- For more information on the ResUsage tables and the types of information each table stores, see Resource Usage Macros and Tables.
The Version screen displays the version numbers of the running PDE and Database System software. Input fields on this screen let you specify different versions of the software to be installed and run the next time the system is restarted.
Example: SCREEN VERSION output
The following is an example of Screen Version command output:
Running PDE: 14.00.00.00 (0) Desired PDE: Running DBS: 14.00.00.00 (1) Desired DBS: Running RSG: (2) Desired RSG: Running TGTW: 14.00.00.00 (3) Desired TGTW: Running TCHN: (4) Desired TCHN: Running TDGSS: 14.00.00.00 (5) Desired TDGSS: Running PDEGPL: 14.00.00.00 Desired PDEGPL:
The Versions screen contains the following control fields.
|Running PDE||The currently running version of Teradata Parallel Database Extensions.|
|Running DBS||The currently running version of Teradata Database.|
|Running TGTW||The currently running versions of Teradata Gateway software.|
|Running TCHN||The currently running version of Teradata Channel software.|
|Running RSG||The currently running version of Teradata Relay Services Gateway software.|
|Running TDGSS||The currently running version of Teradata Database Generic Security Services software.|
|Running PDEGPL||The currently running GNU general public license version of Teradata Parallel Database Extensions.|
|Desired PDE||Changing this field to an installed PDE version causes that version to be run the next time Teradata Database restarts.|
|Desired DBS||Changing this field to an installed Teradata Database version causes that version to be run the next time Teradata Database restarts.|
|Desired TGTW||Changing this field to a an installed TGTW version causes that version to be run the next time Teradata Database restarts.|
|Desired TCHN||Changing this field to an installed TCHN version causes that version to be run the next time Teradata Database restarts.|
|Desired RSG||Changing this field to an installed RSG version causes that version to be run the next time Teradata Database restarts.|
|Desired TDGSS||Changing this field to an installed TDGSS version causes that version to be run the next time Teradata Database restarts.|
|Desired PDEGPL||Indicates an installed PDEGPL version to be run the next time Teradata Database restarts.
Whenever a user sets the Desired PDE to a different version, the CTL tool automatically sets the Desired PDEGPL to the same version.