Details
-
Fix
-
Status: Released (View Workflow)
-
Minor
-
Resolution: Fixed
-
1.12.2
-
None
Description
Current situation
- If Agent is terminated abnormally, the jobscheduler_agent_watchdog.sh watchdog process will execute the ../var_4445/tmp/kill_tasks_after_crash.sh script.
- The kill_tasks_after_crash.sh script is created by Agent when the agent starts a first task. The agent will add an entry to this file for each subsequent task started and remove entries as the tasks are completed.
- If the Agent is not executing any tasks the kill_tasks_after_crash.sh script will either not exist or be empty and if the Agent crashes when tasks are not running the Watchdog will print the message:
2018-03-29 18:18:49,290 +0200 [info] "/opt/sos-berlin.com/jobscheduler_agent/var_4445/tmp/kill_tasks_after_crash.sh" doesn't exist or is empty. Nothing to do."
- This message is confusing and doesn't say that tasks were not running.
Desired Behavior
- The message should only be shown in debug mode.
- The Message should be modified to read, for example:
"/opt/sos-berlin.com/jobscheduler_agent/var_4445/tmp/kill_tasks_after_crash.sh" doesn't exist or is empty. No tasks were running. Nothing to do."