Job

A job is a representation of a batch process. A batch process is usually started due to various system administration related duties, like updating old index, adding new document to the index and databae or creating a new box. A job is immutable. All jobs are cancel-able. Job is a template, not executable unit. Executable unit is a task.

Properties of a job

• Job Id (string, UUID) : Unique id for a job. Unique, generated.
• Job Name (string) : Name of thejob. Non-unique, user assigned.
• Job Type (string) : Type of the job, values could be one of ALL, REFRESH, ERROR. ALL, will create new documents, update old ones from the specified source. REFRESH, will update old documents from the specified source. ERROR, will try to update old documents which failed to get previously.
• Title range (string array) : Document titles, or characters at starting position of titles, eg. ‘ne’ to ‘pq’, defines the range of documents to be used for the job.
• Document source (string) : Source of the documents, can be one of WIKIPEDIA, GUTENBERG, YOUTUBE.
• Scheduled (boolean) : true, if the job is to be run periodically. false, if the job is run once.
• Job frequency (int) : If Scheduled is true, then this number is the number of days between two runs of the job. Default is 15. If Scheduled is false, then this value is ignored.
• Create time (timestamp) : GMT time of job creation.

A few real examples of job :

1. Refresh all Gutenberg documents
2. Retry the failed Wikipedia documents with titles are in range from abc to bcd