Job Templates¶
Parent definition¶
Tag definition¶
<jobTemplates>
<jobTemplate></jobTemplate>
</jobTemplates>
Attributes¶
| Name | Data type and values | Mandatory |
|---|---|---|
| id | Character String type, not empty | Yes |
| jobType | jobType data structure, Table 2 | No |
Table 1: Job Templates attributes.
| Value | Definition |
|---|---|
| mapred_action | Default job Template type. This type will be used as default if no type is defined. |
| wps_client | A job Type which is a connector for external OGC WPS servers. A set of default job configuration is needed (see Table 8). |
Table 2: jobType data structure.
Streaming Executable¶
Description¶
The local path of the executable script called by this job to process the inputs. In the case of jobType=”wps_client” it must be omitted. Supported languages:
- Bash
- Python
- R
Tag definition¶
<jobTemplate>
<streamingExecutable></streamingExecutable>
</jobTemplate
Attributes¶
None
Example¶
<streamingExecutable>/application/mynode/run.sh</streamingExecutable>
Default Parameters¶
Description¶
The default parameters for the job. You can override their attributes and values by using the parameters section of the workflow.
Tag definition¶
<jobTemplate>
<defaultParameters>
<parameter></parameter>
</defaultParameters>
</jobTemplate>
Attributes (for the single parameter)¶
| Name | Data type and values | Mandatory | If “scope = runtime” |
|---|---|---|---|
| id | Character String type, not empty | Yes | Yes |
| title | Character String type, not empty | No | Yes |
| abstract | Character String type, not empty | No | Yes |
| scope | Scope data structure, Table 4 | No | |
| type | Type data structure, Table 5 | No | No |
| target | Query parameter from the description of the source catalogue (mandatory in case of type=”opensearch”) | No | No |
| minOccurs | Minimum number of expected instances of the parameter (WPS). By default is 1 (meaning that it’s mandatory) | No | No |
| maxOccurs | Maximum number of expected instances of the parameter (WPS). By default is 100 | No | No |
Table 3: Default Parameters attributes.
| Value | Definition |
|---|---|
| runtime | The element will be exposed by the OGC WPS interface |
| test | The element will use as value the one defined in the xml |
Table 4: Scope data structure.
| Value | Definition |
|---|---|
| LiteralData | Character String type, not empty (default type if not defined) |
| BoundingBoxData | Coordinates of the Area Of Interest according to the format urn:ogc:def:crs:EPSG:6.6:4326 |
| opensearch | Character String type, not empty. The value is used to query the source catalogue according to the target value. |
Table 5: Type data structure.
Possible values (for the single parameter)¶
In the case of type=”LiteralData” (default “type” if not defined), the parameter element can be defined with:
no values. e.g.:
<parameter id="myparam"/>
a string value. e.g.:
<parameter id="myparam">My param value</parameter>
a list of <options>. e.g.:
<parameter id="myparam"> <options> <option>value1</option> <option>value2</option> </options> </parameter>
optionally it is possible to add a default value to the option list (to be picked from the list of the options):
<parameter id="myparam"> <options> <option>value1</option> <option>value2</option> <default>value1</default> </options> </parameter>
Default Job Configuration¶
Tag definition¶
<jobTemplate>
<defaultJobconf>
<property></property>
</defaultJobconf>
</jobTemplate>
Attributes (for the single property)¶
| Name | Data type and values | Mandatory |
|---|---|---|
| id | Character String type, not empty | Yes |
Table 6: Default property attributes.
Possible IDs and values (for the single property)¶
| id | Data type and values | Definition |
|---|---|---|
| mapred.task.timeout | nonNegativeInteger type | The number of milliseconds before a task will be terminated if it neither reads an input, writes an output, nor updates its status string. |
| ciop.job.max.tasks | nonNegativeInteger type | The number of parallel jobs to be ran, instantiating the job template. This number should fit with the number of available nodes on the cluster. |
Table 7: Possible configuration properties.
In case of jobType=”wps_client” a list of configuration properties has to be defined:¶
| id | Data type and values | Definition |
|---|---|---|
| ogc.wps.access.point | Character String type, not empty | The end point of the remote WPS server. e.g.: http://remoteWPS:8080/wps/WebProcessingService |
| ogc.wps.process.identifier | Character String type, not empty | The name of the process of the remote OGC WPS server. e.g.: com.terradue.wps_oozie.process.OozieAbstractAlgorithm |
| ogc.wps.storeExecuteResponse | true/false | True if the call to OGC WPS has to be asynchronous. |
| ogc.wps.status | true/false | True if the call to OGC WPS has to be asynchronous. |
| ciop.job.max.tasks | nonNegativeInteger type | The number of parallel jobs to be ran, instantiating the job template. This number should fit with the number of available nodes on the cluster. |
Table 8: Mandatory configuration properties in case of jobType = “wps_client”.
