Process flow and batch functionalities (feature request)

nnCron and nnCron LITE discussion

Process flow and batch functionalities (feature request)

Postby RB » Fri, 22 Jul 2005, 15:18

Hey Valery,

Sorry to annoy you but I think I have some nice ideas to enhance nnCron with new predefined functionalities (for Forth-newbies like me).

  • Defining a maximum running time for a task and start some action(s) if it’s not finished in time (i.e. send message, kill application, …).
  • Delay starting of a task within a given time frame if some conditions are set (i.e. another task is still active, a program is running, system-load is to high, user is logged on, …), and if time has passed and task couldn’t be started do something else (similar to above).
  • Add some kind of group identifier to tasks (i.e. networking, reporting, backup, …) which can be checked in conditions (i.e. wait till no other Backup-task is running, …).
  • Make return code of tasks useable as parameters for subsequent tasks (i.e. run only cleanup-tasks if backup-tasks were successful finished).
Background for these absurd requests is my following problem:
I have to generate a lot of OLAP-cubes, reports, … unattended on a machine (Win2003 TerminalServer – I know there are still some problems with).
They have to be done at very different time intervals (last day of month; 2nd, 3rd and 4th day of a month, every Saturday, first working day in quarter, …) which could coincide.
They all work only if a specific oracle-database is online (not in maintenance mode).
And most of them need a lot of time (between 30 minutes and 8 hours).
Additional (and this is very important) they can’t run at the same time - because of system overload, file locking, … (do you know COGNOS? Nice tool, but sometimes …).

Maybe following could be realized:
  • I want to put these tasks on something like a common used task-stack to be processed in batch (serial and not parallel).
  • Every task should have an earliest and latest starting time (which could also be relative to the first one, like in current missed task handling) and a maximum running time (because sometimes COGNOS hangs without cause).
  • Optional I want to define a non-working timeframe (to avoid running tasks in published database maintenance times).
  • Only COGNOS tasks should be grouped together, other tasks (like time synchronization) should work independent.
  • The order of the stack shouldn’t be FIFO or LIFO, it should depend on planned (earliest or latest) starting time (an extra feature would be some kind of priority handling for some tasks).
  • And dependent on the result of the task different actions should be initiated (like an email send to the user of the report on success or otherwise to the administrator/developer)
  • Some other functionalities like
      logging (started, missed, killed, …),
      listing items on the task-stack (current waiting, passed within last day, …),
      changing them (add, remove, modify times, …),
      killing a running task (hanging or misplaced),
    … would also be very useful.

I’m sure this all could already be done with nnCron using Forth, but I’m not a programmer (and don’t want to learn to write such complicated Forth programs, if avoidable). :oops:
I think this features could also be interesting for other nnCron-users and a GUI-based solution would make it much easier to handle.

Ok – you’re right, I’m dreaming about a pink future :finga: - but only dreams put you forward. :D
Best regards, Robert
RB
 
Posts: 4
Joined: Thu, 21 Jul 2005, 16:32

Return to nnCron forum (English)

Who is online

Users browsing this forum: No registered users and 1 guest