How to use a database server for distributed task scheduling?

Question

I have about 100 computers and several workers on each of them. Already connect to the central database to request job parameters.

Now I need to carry out job scheduling for them. One work for one employee takes several minutes, does not require a network connection (except for tasks and reports) and can be performed at any time in any order.

Limitations:

• no work will be done / done twice
• there will be a time when the worker dies.

I thought that I could use a separate table in the database to schedule tasks. How to create and access a job scheduling table?

Answer 1

Divide it into parts:

You have a description of the task, which may have some restrictions on where it is performed, and you have a queue of tasks that need to be performed, generally speaking, if there are no other restrictions that you would expect the tasks to be performed in order, t .e. you take them from the front of the line and add them to the end.

If we run the queue in a separate table, then the tasks will have 3 states:

• Does not work
• Performed
• Completed

- ( n , ), , . , n , n , , .

:

• DateQueued ( )
• DateStarted ( ) -
• AssignedTo

, DateCompleted, , ( ).

, ( "" ) :

• -
• • ( n , )
  • ( , ) - .

, , "-", reset .

.

• ID
• JobID - ,
• StatusID
• DateQueued
• DateStarted
• AssignedToID
• DateCompleted

, ...

Answer 2

, , .

: , . " " .

, , , . , , .

" " . . .

.

" " , . . , (, "" ).

.

- , - " " . , - .

" " . , , , , . , Postgresql , , .