Details
-
Feature
-
Status: Released (View Workflow)
-
Minor
-
Resolution: Fixed
-
1.10
-
None
Description
Feature
- In case of a JobScheduler Agent crash, the Agent start script kills any running tasks.
- The start script starts the Agent with a watchdog process. The watchdog process waits for termination of the Agent. Once the Agent terminated abnormal then the script ./bin/jobscheduler_agent_kill_task.(sh|cmd) is executed by the watchdog process for each running task.
Implementation
- Kill Running Tasks
- The Agent provides the command to execute the kill script for each running task with the file kill_tasks_after_crash_<port>.(sh|cmd) in the logs folder, see
JS-1551. - Before start of an Agent the Agent start script removes an existing file with this name.
- The Agent start script starts the Agent with a watchdog process which waits for termination of the Agent.
- After an abnormal termination of an Agent the watchdog process executes this file should the file be available and should any lines with commands be contained in this file.
- The Agent provides the command to execute the kill script for each running task with the file kill_tasks_after_crash_<port>.(sh|cmd) in the logs folder, see
- Agent Kill Script
- The script is parameterized as explained with
JS-1551 - The Agent kill script will log the information from its parameters to the log file agent_kill_task_<port>.log in the logs directory.
- The script is parameterized as explained with