Hello, I've created flink deployment in native mod...
# troubleshooting
c
Hello, I've created flink deployment in native mode (kubernetes) and tried to submit 10 FlinkSessionJobs, but I see task manager pods are created for each of the flinksessionjob and are getting terminated and many of the submitted flink sessionjobs are stuck in upgrading state. Can someone share the configuration so that task managers are utilized by the other submitted jobs instead of getting terminated
c
Hey @Gyula Fóra I didn't set the property. May I know the unit? Is it milli second?
g
probably millis, this controls how long after a taskmanager is not used it is shut down
you probably also need to configure the number of task slots correctly if you want to share a TM by multiple jobs
c
okay, will try setting timeout property.