scheduler should test for running agents before killing itself
|Assignee:||Alex Norton||% Done:||
Killing the scheduler kills all the running agents. I think we need to add a failsafe such as:
1) if there are running agents, the scheduler needs to tell the user and not do the kill. It should also tell the user to use the -f (see #2 below) if the really want to force the kill.
2) add a new switch, like -f to force the scheduler to die (the current behavior).
3) add a new option to tell the scheduler to not schedule any new jobs and exit when the current ones are done. That is, just finish running the current jobs, email the user who requested this option, then exit.
Updated by Bob Gobeille over 1 year ago
should do a graceful shutdown. That is, it should finish the currently running jobs, email the user that it is shutting down, then shutdown. It should also give the user a message that if they want an immediate shutdown to do an /etc/init.d/fossology kill
Updated by Alex Norton 12 months ago
- Status changed from New to Resolved
This is fixed with svn
Changes the "stop" command from the use interface to wait till all currently executing agents have finished running before exiting the scheduler. Also added a new command: "die" that will cause the scheduler to kill all currently running agents then exit.
Currently the GUI uses the "stop" command and does not offer the "die" command to the users.
Updated by Mary Laser 11 months ago
- Status changed from Resolved to In Progress
Please update the test for this issue http://www.fossology.org/projects/fossology/wiki/Scheduler_Test_Cases#Scheduler-StartStop
Updated by Alex Norton 11 months ago
- Status changed from In Progress to Resolved
added a test case for this. It can be found under src/scheduler/agent_tests/Functional/scheduler_testStop.php