DOCK Blaster:Crash: Difference between revisions

From DISI
Jump to navigation Jump to search
No edit summary
 
m (3 revisions)
 
(2 intermediate revisions by one other user not shown)
Line 1: Line 1:
You think DOCK Blaster may have crashed. We distinguish two kinds of crashes: pipeline errors and platform errors.
You think DOCK Blaster may have crashed. We distinguish two kinds of crashes: pipeline errors and platform errors. This is part of the [[DOCK Blaster:FAQ]].


== Pipeline Error ==
== Pipeline Error ==
Line 14: Line 14:
c. modify the input (if applicable) and try again
c. modify the input (if applicable) and try again
d. write to support at docking.org quoting the job id and explaining the problem only after you have tried the above.
d. write to support at docking.org quoting the job id and explaining the problem only after you have tried the above.
[[Category:Problems]]
[[Category:DOCK Blaster]]
[[Category:User guide]]

Latest revision as of 20:02, 8 October 2012

You think DOCK Blaster may have crashed. We distinguish two kinds of crashes: pipeline errors and platform errors. This is part of the DOCK Blaster:FAQ.

Pipeline Error

A program in the DOCK Blaster pipeline terminated abnormally, perhaps because it was unable to handle the input it was given. This can be handled in one of two ways. Ideally, the job will stop and an intelligible error message will be produced which can guide the user to modifying the input. The system manager will also be informed, as the problem may reflect a "bug" in the software. Sometimes the job may stop without producing a message of any kind.

Platform Error

A job stops because the computer it was running on crashed, or there was some other system-level problem that prevented the job from continuing. We certainly aim to have our computers running 24/7, but crashes cannot be avoided entirely. Often, no error message will be produced in this case. In many cases, the job will be restarted automatically by a supervisor. This process may even go undetected by the end user. However, there are numerous circumstances under which the job cannot be continued for some reason. In this case, resubmitting a new job is indicated.

Recommended Actions

In any case (pipeline error with or without message, or platform error) the recommended actions are as follows:

a. wait an hour to see whether the problem fixes itself. b. resubmit the job, and see whether it fails in the same way c. modify the input (if applicable) and try again d. write to support at docking.org quoting the job id and explaining the problem only after you have tried the above.