why schedule pod always restart?

[sdhzlzhk]

I intalled airflow in my k8s cluster. But the schedule pod always restart when receiving signal 15.
The log info below:
[2021-12-03 11:06:54,214] {scheduler_job.py:555} INFO - Starting the scheduler
[2021-12-03 11:06:54,215] {scheduler_job.py:560} INFO - Processing each file at most -1 times
[2021-12-03 11:06:55,777] {manager.py:163} INFO - Launched DagFileProcessorManager with pid: 17
[2021-12-03 11:06:55,780] {scheduler_job.py:1066} INFO - Resetting orphaned tasks for active dag runs
[2021-12-03 11:06:55,786] {settings.py:52} INFO - Configured default timezone Timezone('UTC')
[2021-12-03 11:11:32,594] {scheduler_job.py:161} INFO - Exiting gracefully upon receiving signal 15
[2021-12-03 11:11:33,596] {process_utils.py:100} INFO - Sending Signals.SIGTERM to GPID 17
[2021-12-03 11:11:39,767] {process_utils.py:66} INFO - Process psutil.Process(pid=497, status='terminated', started='11:11:32') (497) terminated with exit code None
[2021-12-03 11:11:41,195] {process_utils.py:212} INFO - Waiting up to 5 seconds for processes to exit...
[2021-12-03 11:11:41,195] {process_utils.py:66} INFO - Process psutil.Process(pid=498, status='terminated', started='11:11:32') (498) terminated with exit code None
[2021-12-03 11:11:41,222] {process_utils.py:66} INFO - Process psutil.Process(pid=17, status='terminated', exitcode=0, started='11:06:55') (17) terminated with exit code 0
[2021-12-03 11:11:41,223] {process_utils.py:100} INFO - Sending Signals.SIGTERM to GPID 17
[2021-12-03 11:11:41,294] {scheduler_job.py:614} INFO - Exited execute loop

[potiuk]

You need to look at your resources in k8s and k8s logs. Likely something is sending the signal - after finding that there are not enough resources (memory, disk etc ). Memory is most likely.

[sdhzlzhk]

Increase default liveness probe timeout.

 livenessProbe:
    initialDelaySeconds: 60
    timeoutSeconds: 30
    failureThreshold: 5
    periodSeconds: 60

#19003