Job and Cache Management
General - Jobs
This function allows you to view the activity of jobs within a Spark context.
This corresponds to the materialization at the level of the DataChain screens of the execution of the requests.
The materialization of the activity is visible at the level of the DataChain solution banner
Visualization of activity for the DataChain instance and for the current context.
The example shows 4 Jobs in progress.
Jobs are stacked in a queue.
In the example, 3 are Waiting, 1 Running.
An instance can be declined in several work contexts.
(Spark Context)
The activity represented that of a context (the current context)
Jobs and Cache screen description
Enable or disable activity history viewing for this instance and context.
The Cancel all jobs hyperlink cancels all jobs in the queue and the job currently running.
This is a global action on all the jobs.
It is possible to cancel a single Job (in progress or pending).
In this case, position yourself on the line and click on the hypertext link located at the end of the line.
This option enables cache management.
If the option is set to YES then DataChain takes into account the cache if it exists.
The Clear cache link resets the cache for this instance and this context.
Zone that displays all the jobs (by default, the table contains the jobs in progress and the jobs on hold).
If the option See only current jobs is set to No, then all the jobs are displayed, namely:
-
Common jobs
-
Pending jobs
-
The old jobs (History)
What is Display Cache in the DataChain solution
The cache makes it possible to optimize the availability of information from an identical job previously executed by a user. It avoids repetitive execution of jobs in the Spark context. The provision is therefore optimized.
Warning.
The information made available by cache consumption corresponds to the data loaded for display at a specific Date and Time.
An indication is provided to the user when displaying the data.
Example: Data for 3 days 3 hours 10 minutes.
The user will always have a way to run the jobs again without consuming the cache. It will therefore display the current data. A new cache will then be available at the new date and time.