#27 Resource keep-alive

[Anonymous]

Originally created by: martin.r...@gmail.com
Originally owned by: martin.r...@gmail.com

To whom is this feature relevant?
- Resource owners

Please provide a one line description of the feature:
- The MiG server assures that a started resource is alive

How would the new feature work (appear in use)?
Given a keep-alive interval the MiG server checks if a started resource is
still active. If not the resource is restarted by the MiG server.
A MiG resource can die if the machine is restarted or if the MiG resource
scripts for some reason fail.

How would the new feature be implemented?
Through the status command provided when creating/modifying a resource,
the server can check if a resource is still alive, and if not restart it by
using the default (re)start command provided when the resource was
created/modified.

Please provide any additional information below.
The keep-alive feature should be able to handle the gab between job
execution finish and the actual delivery of the output files handled by the
frontend script. As this will be detected as a resource down by the default
resource status script. This can be done by checking the status of the
resource several times before actually calling the (re)start command.

So far this functionality has been handled externally by user keep-alive
scripts, an example of such a script, as well as the input file needed, is
attached.

[Anonymous]

Originally posted by: jonas.ba...@gmail.com

Thanks, but did you forget to actually attach the file? :-)

-Jonas

Cc: martin.rehr

[Anonymous]

Originally posted by: jonas.ba...@gmail.com

(No comment was entered for this change.)

Status: Accepted

[Anonymous]

Originally posted by: martin.r...@gmail.com

Attachments added :)