Help : Job Queue Help
Contents
Description
The database provides access
to software that
requires extended execution time and/or is computationally
intensive. These programs include reading MAGE array design files,
microarray data retrieval and
several others. What often happens with these long running
processes is that a web connection is opened to initiate the process.
The process begins execution still under the control of the (Apache)
web server. This has the effect of needlessly tying up a web
connection. Additionally, after several minutes with little or no output
to the browser the web server may terminate the process. Further, if there are
many users running computationally intensive applications simultaneously,
there may be negative effects on overall system performance.
The job queue and job manager were
invented to alleviate these difficulties.
The job manager may run n jobs at a time. The database staff monitors system
performance and may adjust this parameter depending on system load and
types of computation currently active.
Using the Job Queue
Jobs are typically entered
into the job queue via a web page. The
results are usually deposited in the user's repository and the user is
then notified via email when the job completes. Applications that
support the job queue often provide a check box in the web form such as:
There are typically options to name the resulting data file in your repository and to add some meaningful comment. Please note that you must name your job in order for the job to run in the job queue. After your job has been placed in the job queue, you should see a page similar to the following indicating that your job was successfully placed in the job queue:
Accessing Results
When the Job Manager selects a job to run, a log file is created and the user is notified via email that the job has started. A URL to the job log is included in the email. Another email is sent by the Job Manager when the job is complete, again including the URL to the log file. The job log will be deleted after approximately 36 hours. Thus, if the job log contains information or data that is of interest to the user, it is the user's responsibility to save this information. A typical job log is as follows: