Details
-
Fix
-
Status: Released (View Workflow)
-
Major
-
Resolution: Fixed
-
1.10.5, 1.10.6
Description
Current situation
When running with HTTPS and no task is being created for more then 10 minutes no more task can be started. In the log there is a message
2016-11-16 12:26:32.984+0100 [WARN] SCHEDULER-488 This remote JobScheduler 'Agent(0,https://192.11.0.116,None)' is not accessible: spray.can.Http$ConnectionAttemptFailedException: Connection attempt to 192.11.0.116:443 failed
2016-11-16 12:26:33.010+0100 [WARN] SCHEDULER-489 No remote JobScheduler is accessible. Waiting before trying again
This does not happen when at least every 5 minutes a task will be started on the agent.
Desired Behaviour
The connection should not be lost even when no task has been started for a long time > 10 minutes
Workaround
Set the path to the Java Keystore
The recommended workaround is to set the path to the Java Keystore with the file $SCHEDULER_DATA/config/factory.ini. Please add:
"-Djavax.net.ssl.trustStore="${SCHEDULER_DATA}/config/agent-https.jks" to the options. Then restart JobScheduler Master.
[java]
options ="-Djavax.net.ssl.trustStore='${SCHEDULER_DATA}/config/agent-https.jks' -Dlog4j.configuration='file:///${SCHEDULER_HOME}/lib/log4j.properties'"
Periodically run a job
An alternative workaround is to use a periodically running job like this:
<job>
<script language="shell">
<![CDATA[
exit 0
]]>
</script>
<run_time >
<weekdays >
<day day="1 2 3 4 5 6 7">
<period absolute_repeat="00:03" begin="00:00" end="24:00"/>
</day>
</weekdays>
</run_time>
</job>
Attachments
Issue Links
- mentioned in
-
Page Loading...