Data Persistence Options:

 

Option

Argument

Description

-p

--port

port-spec

Listen on the specified ports.

Port-spec is either plaintext-port:encrypted-port, or just a plaintext port number.

One of these can be disabled by specifying OFF instead of a port number. Disabling both is not allowed. The default plaintext port is 5051. The default encrypted port is 8051.

Note, this only specifies port numbers. Encrypted communications are disabled by default, they can be enabled by -k and -K options.

-g

--gateway

config-file

Run as a gateway server using the specified gateway configuration file. Gateway servers and gateway configuration files are explained in the section, Gateway Servers and Clustering.

-P

--persist

socket-option

Use persistent socket connections. socket-option is one of I = Incoming only, O = Outgoing only, B (default) = both, N = neither.

-u

--ip-in

ip-mode

Set the incoming IP protocols the server accepts. ip-mode can contain one of the following protocols:

4 - IPv4 only

6 - IPv6 only

M - Use both protocols. The default value is M.

-U

--ip-out

ip-mode

Set the outgoing IP protocols the server uses. This controls what protocol this server uses when creating connections to a remote server. ip-mode can contain one of the following protocols:

4 - IPv4 only. The default value is '4', use only the IPv4 protocol for outgoing connections.

6 - IPv6 only

M - Use both protocols.

ip-address...

 

Accept connections from the specified IP addresses as well as "localhost". By default, the server only accepts connections from "localhost" (127.0.0.1 or [::1] depending on the IP mode selected).

The authentication list consists of localhost (127.0.0.1 or [::1] depending on the IP mode selected) along with any addresses given in this address list. Only hosts in the authentication list can connect to the server. Addresses in the list may be any host name that can be resolved to an IP address. If IPv4 is enabled (it is enabled unless -u 6 was specified), addresses in this list may also be valid IPv4 format IP addresses. Wild card and subnet mask lengths may be used. For example: 129.48.32.* 192.168.*.* 129.48.34.0/24. Note that the entries may need protection from being expanded by the shell or command interpreter. If IPv6 is enabled (the default on all platforms except Windows platforms predating Windows Vista, and if -u 6 or -u M are specified), IPv6 format IP addresses may be given. The standard text formats as defined by RFC-5952 are supported, including subnet masks. As with IPv4 addresses, these must be protected from expansion by the shell or command interpreter.

Encrypted Communications Options:

Option

Argument

Description

-K

--private-key

filename

Private key file-name.

If a single file contains both the private key and the public key, then either -k or -K can be omitted. If both are omitted, encrypted communications are disabled.

-k

--public-key

filename

Certificate file-name.

-Z

--trusted-store

trust-store

SSL trust store file name or directory name.

If this option is omitted, the system-default trust store is used. To specify a system certificate store on Windows, prefix the store name with wincs: The standard system stores are ROOT, MY, SPC, and CA. The default is wincs:ROOT.

Search Options:

Option

Argument

Description

-F

--prefilter

GIP | SORT | PSI

Specify the default prefilter. If this argument is not given the default prefilter is GIP. The default prefilter can be overridden at the time the table is created with an explicit prefilter selection.

-T

--thes

sub-limit

This specifies a limit on the number of different words or phrases in a query string to which thesaurus substitutions are applied. Substitutions beyond the limit are ignored. There can be a large performance penalty for very large numbers of substitutions, this limits that penalty. The default is 40, which is more than enough for most applications.

-J

--joins

join-limits

 

Sets limits for joined searches.

Preferred format for join-limits:

[single-parent-rec]:[all-parent-recs]:[max-querylets]

single-parent-rec is the maximum number of child record combinations for a single parent record that is found by the GIP prefilter
all-parent-recs is the maximum number of such combinations for all parent records returned by the GIP prefilter
max-querylets is the maximum number of leaf querylets allowed in any query
Note: All of the above three are optional, but the two colons are required. For example, if single-parent-rec and all-parent-recs are not applicable, and max-querylets value is 2000, the format is -J ::2000

Alternative legacy format for join-limits:

- J single-parent-rec[:all-parent-recs]

Note: If only the single-parent-rec limit is specified, then all-parent-recs is set to be equal to single-parent-rec and max-querylets is left at the default. If only the single-parent-rec is used in this format, then the colon must be omitted.

The alternative legacy format provides backward compatibility for existing applications, but the preferred format is more flexible and allows setting the maximum number of querylets.

For more information, see the “Single Parent Search” and “Matching Compound Records” sections in TIBCO Patterns Concepts Guide.

-C

--gpu

 

Enables Graphics Processing Unit (GPU) harnessing for each table and query. For information about GIP-GPU, see the "Prefilters" section in TIBCO Patterns Concepts Guide.

Note: -C is not supported on gateways, that is, starting a server with -g config-file -C is not a supported configuration. For information, see Gateway Servers and Clustering.

LIST

Lists available GPU devices and exits without starting the server. It indicates whether it meets recommendations of TIBCO for each device.

Using non-recommended devices can degrade application performance.

RECOMMENDED-ONLY

Starts the server with recommended devices harnessed. Usable, but non-recommended devices are not harnessed and cannot be requested by applications.

DEFAULT

Starts the server using GPU devices according to TIBCO recommendations. By default, recommended devices are harnessed. Usable, but non-recommended devices are harnessed only if requested by the application.

ALL

Starts the server with all usable GPU devices harnessed.

device-specification

Starts the server using specific GPU harnessing.

device-specification is a comma separated list of one or more pairs of device-id,flag[, device-id,flag]*

device-id: Device IDs can be obtained by using -C LIST.
flag: The flag value can be D or O.
D for a default device. Default devices are harnessed to tables if not specified by the application.
O for an optional device. Optional devices are harnessed to tables only if requested by the application.

Data Persistence Options:

Option

Argument

Description

-R

--restore-dir

restore-dir

Enable checkpoint and restore using the indicated directory. The directory must already exist. If this argument is not given, the checkpoint feature is disabled.

-A

--auto-restore

 

Auto restore. Upon start up, restore all checkpointed tables in the restore directory. This is ignored if the -R option is not specified.

-a

--restore-from

common-dir

Restore from a common directory. Upon start-up all checkpointed tables in the indicated directory are restored. Checkpointing is not enabled by this option, although it may be used with the -R and -A options.

-B
--durable-data
  Enable the Durable Data feature. This requires the -R option. Incompatible with the -a option.

Logging Options:

Option

Argument

Description

-c

--con-log

console-log

Write error messages to the specified file. By default, the server sends messages to the system log on UNIX and the application event log on Microsoft Windows.

-l

--query-log

query-log

Log incoming queries to the specified file.

-v

--verbose

Log additional data to the console log. Each connection attempt is logged. If a connection attempt is rejected, the reason is logged. Do not use this option in production as the console log grows rapidly, until the file system is full.

-D

--socket-log

socket-log

Logs all messages sent and received by the server to the indicated file. As it produces a massive file, use it only at the request of TIBCO Support.

Information Options:

Option

Argument

Description

-M

--model-info

model-file

Display header data for the specified Learn Model file, and then exit.

-V

--version

 

This causes the TIBCO Patterns server to print out its version information and exit.

-d

--debug

 

Start in debug mode. The server does not detach itself from the controlling terminal, nor does it put itself into the background.

-h or -?

--help

 

Help. Print a usage message and exit.

Other Server Configuration Options:

Option

Argument

Description

-G

--temp-dir

temp-storage-dir

This directory is used by TIBCO Patterns for temporary files. The file system containing this directory must have sufficient free space to store any model files transmitted through the gateway.

-t

--threads

max-threads

Spawn no more than the specified number of command processing threads. The default is 4.

-r

--pid

pid-file

 

Write the server's process ID to the specified file. The file is deleted when the server terminates.

-H

-hawk

"service, network, daemon"-

Enable the Hawk interface. If this argument is given, the server registers with TIBCO Patterns Hawk as a micro-agent, enabling TIBCO Hawk to monitor the state of the server. service, network, and daemon are the transport parameters as defined for TIBCO Hawk. For more information, see TIBCO Hawk Interface.

-m

--memory

memory-cap

Set a limit for the total amount of space used for all database tables in the server. The space is given in KB. The minimum allowed size is 1024. When the total memory used for all database tables in the process reaches the limit specified, any attempt to add additional records or tables is disallowed. If this argument is not given, no check is made. If enough memory cannot be allocated for an additional record or table, the process terminates. Thus this argument can be used to protect the server from aborting due to insufficient memory resources or to cap its memory use on shared systems.

-L

--max-loads

max-loads

Set the limit for the number of parallel loads in a multitable restore. By default, the Auto restore on start-up loads up to maxthreads and three tables in parallel to reduce load time where maxthreads is as defined in the -t option. During command processing, parallel loads in multi-table restores are turned off by default as they may interfere with other command processing. This can be controlled through this option. Setting max-loads to a positive value allows up to that number of tables to be loaded in parallel during command processing. Setting it to zero removes all restrictions and loads all tables in parallel. Setting it to a negative value disallows parallel loading in multitable restores during the command processing (same as the default behavior). For the initial auto-restore, the limit is always the maximum of maxthreads and three or max-loads.

-I

--idle-tran

idle-timeout

Controls the auto-timeout of idle transactions. Idle-timeout has the following format: idle-time[:action] where idle-time specifies the time in seconds. A transaction must sit idle before the action is applied. An action can contain one of the following arguments:

A – Abort the transaction

E – Abort the transaction if it has any Errors, commit if it is otherwise

C – forcibly Commit the transaction

N – perform No action, this turns off idle transaction monitoring. If this argument is not specified, idle transaction monitoring is turned off. If an action is not specified, it defaults to A.

-E

--loadable-data-dir


loadable-data directory Specifies the directory for server-side data loads. The server receives a notification from the application to load a server-side data file following which, only server-side data files are stored in this directory. The default directory path is workingdirectory/loadable_data. In case of incorrect loadatable_data_dir, TIBCO Patterns shows a WARN message. You can also provide relative path-. inside loadable_data_dir.

UNIX only: on startup, the server detaches itself from the controlling terminal and redirects all output to the console and log files as specified in the command line arguments. This is the standard behavior for background "daemon" processes on UNIX platforms. The -d option disables this behavior.