Project

General

Profile

Actions

action #1711

closed

cleanup jobs of dead workers

Added by lnussel about 10 years ago. Updated about 10 years ago.

Status:
Resolved
Priority:
Normal
Assignee:
Category:
-
Target version:
Start date:
2014-02-24
Due date:
% Done:

100%

Estimated time:
8.00 h

Description

sometimes workers may crash, leaving the job in e.g. running state. since workers ping the webui every second, updating their table row there should be a job that periodically checks for workers that are no longer active.

Actions #1

Updated by lnussel about 10 years ago

  • Estimated time set to 8.00 h
Actions #2

Updated by lnussel about 10 years ago

  • Assignee set to mlin7442
  • Target version set to Sprint 03
Actions #3

Updated by lnussel about 10 years ago

Max, I think the detection of the dead worker could be more easily done with a database query, isn't it? Ie select running jobs, join with workers and select the rows where the t_updated is older than some threshold.
You cannot use job_cancel as that one inserts a command for the worker if a worker is assigned. You need to use job_set_done and job_duplicate instead.

Actions #4

Updated by mlin7442 about 10 years ago

  • Status changed from New to In Progress
  • % Done changed from 0 to 100

ok, $threshold, job_set_done and job_duplicate are there now. I'll polish it later with a database query.

Actions #5

Updated by mlin7442 about 10 years ago

  • Status changed from In Progress to Resolved

done

Actions

Also available in: Atom PDF