TSTool / Running TSTool in Various Modes
- Overview
- Running TSTool in Interactive Mode
- Running TSTool in Limited Interactive Mode
- Running TSTool in Command Line Batch Mode
- Running TSTool as a Scheduled Task
- Running TSTool as a Web Server
- Running TSTool as a Batch Server
- Running TSTool as a REST Web Server
- Running TSTool as a Web Browser Applet
- Substitute Datastore Examples
Overview
TSTool is commonly run as an interactive desktop application using the graphical user interface (GUI). However, the TSTool design allows the software to be run in various modes, as described in this document:
- Desktop application
- Interactive graphical user interface
- Limited graphical user interface (e.g., no main GUI but graph windows display in support of other applications)
- Command line batch mode (e.g., Windows command shell or Linux shell script)
- Scheduled task (e.g., Windows scheduler or Linux cron)
- Web server application (e.g., run headless processing in support of website)
- Batch server to process command files in a hot folder (
-batchServer
mode) - REST web server (e.g., run TSTool as a stand-alone REST web server)
- Web browser applet
The following sections focus on each run mode for various operating systems. In addition to these run modes, the TSTool installation configuration may involve the software being installed on a local computer for one user or a server that provides access to multiple desktop computers and users. It is possible for any of the run modes to be configured on a desktop computer, given the proper supporting software and expertise; however, it is likely that server-oriented run modes will be configured on servers. The following sections describe how to run TSTool in various modes. It is assumed that TSTool has been properly installed and configured based on information in other appendices.
Running TSTool in Interactive Mode
Running TSTool in interactive mode using the graphical user interface (GUI) is the primary run mode for most users, in particular when editing and running command workflows. The Getting Started chapter and other chapters in the TSTool documentation focus on this run mode and are not discussed further here.
Often as TSTool processing becomes integrated with an organization’s business processes, it is appropriate to run TSTool command files in batch mode for efficiency or as a scheduled task to support other tasks. The following sections describe how such processing can be implemented. Because the command files are the same regardless of whether TSTool is run in interactive or batch mode, the interactive tool can be used to modify and troubleshoot command files, as appropriate.
Running TSTool in Limited Interactive Mode
It may be useful to run TSTool in support of other applications.
For example, TSTool has been integrated with Geographic Information System (GIS)
desktop applications where clicking on a map feature results in running
TSTool to process data to produce an interactive graph.
In this case, TSTool is run in batch mode (see following sections) and
the –nomaingui
command line parameter is used to indicate the special run mode.
The user can interact with the displayed visualization window as if the window had been created from the normal GUI,
including changing between graph, table, and summary views, changing view properties, and printing.
Potential issues with this approach are:
- Error-handling and feedback to the user are more difficult because TSTool
is not fully integrated with the calling software.
The
StartLog
command can be used to ensure that a log file is created in a location that is integrated with the overall system. - There may be a slight delay displaying the view window because TSTool starts each time.
To improve performance, a specific configuration for TSTool may be required (see the
–config
command line parameter). For example, disable all datastores that are not necessary for the application because the time required to connect to databases will slow down TSTool startup. Another option is to point to data products (time series graphs, etc.) provided by a web server. This involves implementing TSTool on a web server, which his discussed in a section below; however, such an approach requires implementing more technology components. - It can be confusing to specify the location of configuration information and input files. Using absolute paths limits flexibility in changing the configuration. However, relative paths must be “anchored” to a starting location. For example, TSTool assumes an initial working directory depending on software startup and then sets the working directory to the location of the command file. It generally is best to utilize the top-level application to provide the working directory, and then pass the information to TSTool dynamically if possible.
Running TSTool in Command Line Batch Mode
“Command line batch mode” refers to running TSTool from a command prompt, such as the Windows command shell or Linux command line. The TSTool GUI is useful because it provides immediate feedback on whether commands are properly defined and input data are available. However, once a command file has been defined, it may be appropriate to process the commands in an automated “headless mode”, without the GUI. Run TSTool on the command line as follows on Windows:
[PathToTSTool]tstool –commands [PathToCommandFile]commands.TSTool [TSTool parameters]
Run TSTool on the command line as follows on Linux (note the double dash parameter,
which separates the tstool
script parameters from TSTool program parameters):
[PathToTSTool]tstool -- –commands [PathToCommandFile]commands.TSTool [TSTool parameters]
See the following sections for specific issues that may occur on different operating systems. The following are potential issues on any operating system:
- TSTool relies on a “working directory”, which is the root location to which relative paths are referenced.
Normally the working directory is the location of the command file once opened.
However, when running TSTool in batch mode, there are a number of folders involved:
the location of the TSTool executable, the location from which TSTool is run, and the location of the command file.
These multiple locations can make it difficult to troubleshoot.
One option is to use absolute paths in the TSTool command line for the executable
and the command file so it is very clear where the TSTool executable and command file are located.
These paths can be coded into a batch file (
*.bat
on Windows, or shell script on Linux). - TSTool can be run in headless mode and still create graph windows
(e.g., when TSTool supports the functionality of another application).
Make sure to evaluate whether the
–nomaingui
parameter is needed in addition to–commands
(see the previous section). - The TSTool default log file under the logs folder of the installation is used when TSTool first starts up.
If a
StartLog
command is used in the command file, it will be used when the command file is run. Refer to the appropriate log file when troubleshooting.
TSTool Command Line Parameters
The following table lists the TSTool command line parameters.
If running on Linux using the tstool
script,
the parameters should be specified after --
parameter.
TSTool Command Line Parameters
Command Line Parameter | Description | Default |
---|---|---|
CommandFile |
The command file can be specified without –commands to start up the TSTool GUI and load the command file. This behavior occurs when selecting a *.TSTool file in Windows Explorer. See also -runcommandsonload. |
|
Property=Value |
Define TSTool global application property (similar to the configuration file). These properties typically are used during development to test specific features. | |
Property==Value |
Define global processor properties (similar to the SetProperty command). These properties are initialized for each command file run and are useful for configuring an environment when starting TSTool from a script, such as running in a Linux environment. |
|
-batchServer --batchServer |
Run TSTool in headless mode to process command files that are copied into the hot folder (see -batchServerHotFolder ). TSTool will exit when a command file named stop is placed in the hot folder. |
|
-batchServerHotFolder Folder --batchServerHotFolder Folder |
Specify the hot folder to use with the -batchServer parameter. |
|
-batchTimeout TimeoutSeconds --batchTimeout TimeoutSeconds |
Timeout in seconds. If TSTool command processing takes longer, TSTool will exit. This may be appropriate as a final fallback if other timeout controls do not work. Warning – this will leave processing incomplete. May be removed in future. | Do not time out. |
-commands CommandFile --commands CommandFile |
Specify the name of the command file to run in batch mode. TSTool will process the commands and exit. | Start TSTool in interactive mode. |
-config ConfigFile --config ConfigFile |
Start TSTool using the specified configuration file. This is useful in software test environments. | Start TSTool using the system/TSTool.cfg file under the software installation. |
-dTerm,Log |
Specify the debug level (0+) for terminal and log file debug messages. This is useful for printing troubleshooting messages at startup. One or both values can be specified. | No debug messages are generated. |
--datastore-substitute= DatastoreNameToUse ,DatastoreNameInCommands |
Substitute a datastore name to use for a datastore name used in commands. For example, all automated tests may use a datastore name datastore-test in command files and a production system's DatastoreNameToUse can be specified as a substitute using --datastore-substitute=DatastoreNameToUse,datastore-test . See the Substitute Datastore Examples documentation. |
No substitutes - the datastore name in command files must match a configured datastore name. |
-home InstallFolder --home InstallFolder |
Specify the install folder. This parameter is used internally by the TSTool launcher (Launch4J) and in the software development environment to specify the TSTool home, in order to locate other files. | Should always be set by the TSTool launcher. |
-httpServer --httpServer |
Run TSTool in headless mode to act as an HTTP server. Experimental | |
-nodiscovery --nodiscovery |
Do not run discovery on commands as they are loaded. This can be used for large command files that will not be edited interactively, in order to decrease load time. | Run discovery on commands as they are loaded. |
-nomaingui --nomaingui |
Do not show the TSTool main GUI when running commands, but do allow graph windows to be displayed. Closing the last visible graph window will close the TSTool application. | If –commands is specified, do not show any windows. If -nomaingui also is specified, allow graph windows to show. |
-runcommandsonload --runcommandsonload |
Load and run the commands, used to start the GUI with a specific command file. | |
-server --server |
Start TSTool as a REST restlet server (under development). | |
-test --test |
Run TSTool in test mode, for developers. | |
-ui-titlemod=TitleMod |
Set the TSTool main UI window title to TSTool - TitleMod - ..... This is useful for helping users identify custom TSTool configurations, for example when running a test suite that requires a Linux sudo login. |
Running TSTool in Command Line Batch Mode – Windows
To run TSTool in batch mode on Windows, first open a command prompt application (cmd
).
Then run the TSTool program as follows.
[PathToTSTool]tstool –commands [PathToCommandFile]commands.TSTool [TSTool parameters]
Issues specific to Windows include:
- If the folder containing the TSTool executable (
tstool.exe
) is in thePATH
environment variable, then no leading path is needed. However, because different versions of TSTool may be installed on the system, specifying the leading path to the executable is often be appropriate. - Entering the TSTool command line in a Command Prompt window causes the prompt to be immediately returned,
even though TSTool is still running.
This can be disconcerting in particular because it may be difficult
to know when TSTool has finished processing the command file.
Placing the command in a batch file (
*.bat
or*.cmd
) can help.
Running TSTool in Command Line Batch Mode – Linux
To run TSTool in batch mode on Linux, first open a command shell window. Then run the TSTool program as follows.
[PathToTSTool]tstool -- –commands [PathToCommandFile]commands.TSTool [TSTool parameters]
Issues specific to Linux include:
- If the folder containing the TSTool executable (
tstool
script) is in thePATH
environment variable, then no leading path is needed. However, because different versions of TSTool may be installed on the system or may not be in thePATH
, it may be necessary to specify the leading path to thetstool
executable. It may be useful to use symbolic links to provide access to the current version as tstool, for example:
/usr/bin/tstool -> /opt/tstool-13.03.00/bin/tstool
The linux tstool
script provides the following command parameters.
The --help
parameter should be used to print the current usage.
tstool
Script Command Parameters
Parameter | Description | Default |
---|---|---|
--addpath=path |
Exploratory - add a folder/file to the class path. | The TSTool software install folder /plugins folder and user's .tstool/NN/plugins folders are used to load plugin jar files. |
--dev |
Run using software development files in the Eclipse development environment, used with automated testing. This allows TSTool to be run with development files rather than building and running the installer. | Use production files. |
--headless |
Run in headless mode, used on Linux when running batch and unattended processes. The -Djava.awt.headless=true parameter is used to run Java. |
|
-h --help |
Print the program usage. The help for this script is printed and then TSTool program is run with --help . TSTool command parameters are documented in online documentation. |
|
--java-xmx=NNNNm |
Set the Java program -Xmx option to indicate the maximum memory to use, for example --java-xmx=1024m for 1024 MB. |
1024m |
-v --version |
Print the program version. The version for this script is printed and then TSTool program is run with --version . |
|
-- |
All command parameters after double dash are ignored by the tstool script and are passed to the TSTool software. |
Running TSTool in Command Line Batch Mode – Mac
TSTool execution on Mac computers is similar to Linux.
Running TSTool as a Scheduled Task
Examples of running TSTool as a scheduled task include automatically processing real-time data, and creating data products prior to viewing (rather than creating upon request). The following sections describe how to configure TSTool to run as a scheduled task.
Running TSTool as a Scheduled Task – Windows
Windows has a task scheduling function, which can be used to run TSTool. These instructions are for Windows 7. However, all versions of Windows have similar features. The following references can be consulted for other Windows versions:
To define a task, first go to the System and Security subpanel of the Control Panel, then click on Schedule tasks.
Running TSTool – Schedule as a Task (see also the full-size image)
A screen showing currently scheduled tasks will appear. Click on the Action menu in the upper left, and the Create Task... option.
Running TSTool – Creating a Task in the Task Scheduler (see also the full-size image)
A tabbed dialog will appear. Go to the Actions tab, and click New... to create a new task.
Running TSTool – Create a New Task
Specify the program to run by using the Browse... button to pick the TSTool executable program, as shown below. The example shown on the right is a typical location for the install (specify an appropriate version number).
Running TSTool – Specify Program to Run for Task (see also the full-size image)
Specifying the absolute path to the command (starting at the drive) is recommended.
Also specify that TSTool should run without the user interface and specify command file to
use (-commands CommandFileName
), as shown in the following figure:
Running TSTool – Specify Program and Command Line Parameters to Run for Task
TSTool log files will by default be in the logs folder under the TSTool installation.
Use the StartLog
command to control the log file location.
To actually run the task, define a trigger in the Triggers tab of the Create Task dialog (press New... to create a trigger):
Running TSTool – Specify Trigger to Indicate Run Task Schedule
Any number of triggers can be chosen, and it can be run at almost any interval, from every five minutes upward. Conditions that will prevent the program from running can also be chosen (Conditions tab). Output files will be overwritten on each run unless the filename is dynamic, such as using the date/time.
Running TSTool as a Scheduled Task – Linux
These instructions apply to variants of Linux and UNIX.
Cron is a background system process that checks every minute to see if any process should be run.
To schedule a run, the crontab
command is used.
It creates or edits an entry in the user’s crontab
configuration file.
Each user can only have one crontab
in most variants of Unix (Debian is slightly different; check the man page for more information).
The command crontab -e
will edit an existing crontab
or create a new one if none exists.
If this is the first time a crontab
has been created on the system,
the command may present a list of editors to choose from. The choice is not important to the functioning of the command.
The syntax of a crontab entry is as follows:
minute hour day_of_month month day_of_week command
A *
in any of the first five fields will run the process at every instance,
and a range (e.g. 1-5
or 5,10,15
) can be entered into any of the first five fields.
For example, to run once an hour, the entry would be as follows:
0 * * * * (command)
To run TSTool as a cron process, command should be specified as follows:
/PathToInstallation/bin/tstool -- --commands PathToCommandFile
The paths should be specified as absolute path because cron does not necessarily have a home directory.
The log files will by default be in the logs folder under the TSTool installation.
The StartLog
command can be used to configure the log file location.
Output files will be overwritten at each execution of the command unless the names are determined dynamically.
For more information, see:
- man cron - information about cron options and system variations
- man crontab - information about options and system variations
- man 5 crontab - information about syntax and limitations
Running TSTool as a Web Server
If TSTool is started with the –-httpServer
option it will run as a web server on port 8000.
This functionality is experimental and requires additional development.
This may replace or be combined with the experimental REST servlet describe below.
Running TSTool as a Batch Server
If TSTool is started with the –-batchServer
option it will run continuously in headless mode.
Use the -batchServerHotFolder Folder
command line parameter to specify the hot folder for command files.
This is an experimental feature.
- Each command file that is copied into the folder will be executed and then removed.
If necessary, use the
StartLog
command to create a log file in a different folder to troubleshoot. - Create a file named stop in the hot folder to cause TSTool to stop processing and exit.
- In the future additional features will be added.
Running TSTool as a REST Web Server
It is possible to use the Restlet package to implement a REST API wrapper around TSTool’s batch features. Consequently, TSTool can be started as a REST server, listen for requests, process data, and return resources. Preliminary work has been done to implement this capability but much more effort is needed to provide production features.
Running TSTool as a Web Browser Applet
It is possible to integrate TSTool with a web application to run TSTool as an applet. In this case, the web browser runs the Java TSTool application within a Java Runtime Environment within the web browser. This approach was evaluated early in TSTool development as is still enabled in some form, although it has not been used for years. The main challenges with applets are the security issues of accessing local files from the applet. Because TSTool is used for heavy processing using local data, the applet approach was not suitable. Enabling TSTool as an applet, using remote data (such as web services) is possible but will require additional development and testing.
Substitute Datastore Examples
Datastore names can be substituted to facilitate implementing automated processes. This is used in automated testing.
The following examples illustrate cases that have been used.
In all cases, the --datastore-substitute=DatastoreNameToUse,DataStoreNameInCommands
syntax expects that the first datastore is enabled and working.
The second datastore does not need to be defined or enabled given that it is a substitute name that will point to the original datastore name.
The TSTool main interface can interact with either datastore as long as a datastore is defined, enabled, and working.
Automated Web Services Tests Use Specific Datastore Name and Subtitute Uses a General Datastore Name
Command files can be written to use a specific datastore name because tests were written for a specific system, for example matching the operational system's database used by web services. This allows read-only (non-destructive) testing command files to be run on a production system to confirm functionality and performance. A datastore substitute can be used to point the datastore used with testing to an appropriate server-side configuration, such as a copy of a production system's database. Using a system-specific datastore name helps avoid confusion because specific tests are clearly identified for a specific system.
To execute the tests in the development environment,
the datastore used for the testing framework must be used.
Web services software typically points to one database and consequently,
the database on the web services server used in the testing environment must be swapped to match tests.
The TSTool #@require
comment can be used to ensure that the appropriate database is used.
The following approach is used to configure datastores for testing:
- Datastore configuration 1:
- The datastore is configured and enabled for
datastore1
(a more specific name would be used) to point to a production system:- allows read-only command file to be tested on the production system
- the datastore matches the production system because tests are specific to the system
- station identifiers and other data are specific to the system and tests
- Test command files use specific datastore name
datastore1
, indicating tests that are specific to the system. - The testing framework uses a general datastore name pointing to the development web services, for example
wsname-osname-dev
, wherewsname
andosname
are specific to the web services and operating system and indicate a web services datastore on a specific operating system version, in development environment. For example, the web services back-end may use Java Hibernate to point to a specific database, and the URL on the testing machine is the same regardless of which back-end database is used.
- The datastore is configured and enabled for
- Datastore configuration N:
- Repeat the above.
- Datastores are configured for each production system.
- Testing framework uses a generic datastore name pointing to the development web services, shared among all datastores.
The overall testing approach is then:
- Configure datastores to point to production systems (see above).
- Create test command files that use the production system datastore names (
datastore1
). - Configure a datastore to point to the testing server using a general name (e.g.,
wsname-osname-dev
), wherewsname
andosname
are specific to the web services and operating system. The datastore name will be reused. - Run TSTool to run tests:
- Use command parameters such as
--datastore-substitute=wsname-osname-dev,datastore1
and to cause the datastoredatastore1
found in command files to be replaced with substututewsname-osname-dev
(the datastore name ot use) when run. A shell script is typically implemented to specify these parameters in the development environment. Datastore substitutes are typically NOT used in a production environment, although they could be used to run read-only tests on a production system to evaluate functionality and performance. - Implement a mechanism to reset the web service server-side database to point to a system-specific database,
available through web services via the general datastore name (
wsname-osname-dev
). For example, a special web service endpoint can be implemented to change the server-side database and restart web services, available only in development mode. - Run the test suite for a specific system database, using a script to run TSTool and command files from step 2.
- Use command parameters such as
The TSTool main interface behavior is as follows. If a --datastore-substitute
parameter is used:
- if the
datastore1
datastore is used in the TSTool interface, thewsname-osname-dev
datastore will be used instead - the
wsname-osname-dev
datastore can also be used directly
Automated Database Tests Use a General Datastore Name and Substitute Uses a Specific Datastore Name
Automated tests can be written using a generic datastore name, for example when the tests are intended to be run on any database, including an empty database. This is appropriate when the tests do not depend on the contents of a database. In this case, the data in the database are synthetic and should not conflict with any production database.
The following approach is used to configure datastores for testing:
- Datastore configuration for general (empty) database using a specific name:
- Use a general datastore name such as
system-test-general
, wheresystem
is a specific system/product name. This datastore name can be used generally for testing or a system-specific datastore could be used (see below) - An empty testing database will be used to run tests.
- The testing environment should mimic a production system as much as possible.
- Use a general datastore name such as
- (Optional) - Datastore configuration for system database:
- Configure a datastore for a local production database
- This will allow running tests on a copy of a production database, such as restored from a database backup.
- The testing environment should mimic a production system as much as possible.
- Datastore substitute:
- See the following approach.
The following approach can be implemented to allow flexibility to run the tests on any database, including the generic empty database and system-specific databases.
- Configure a datastore to point to the testing server using a general name (e.g.,
system-test-general
), wheresystem
is a specific system/product name. - Create test command files that use a general datastore name (e.g.,
system-test
), wheresystem
is a specific system/product name. This name is not the same as the datastore name configured above in order to provide flexibility for substitution. These tests will work on any database because the data are synthetic. - Run TSTool to run tests:
- Use command parameters such as
--datastore-substitute=system-test-general,system-test
to cause the datastoresystem-test
found in command files to be replaced with substututesystem-test-general
datastore name when run. A script is typically implemented to specify command line parameters in the development or testing environment. - Run the general test suite, using the command files from step 2.
- If appropriate, use a production system database backup for testing,
with the first datastore name used with
--datastore-substitute
corresponding to the databae.
The TSTool main interface behavior is as follows. If a --datastore-substitute
parameter is used:
- if the
system-test
datastore is used in the TSTool interface, thesystem-test-general
datastore will be used instead - the
system-test
datastore will not be listed in TSTool because it is not configured as a datastore (it is a placeholder in command files)