Is Flink standalone cluster manual restart graceful wrt to job state/processing offset?

90 views Asked by At

When we manually stop a standalone cluster(stop-cluster.sh) and then start it again(start-cluster.sh), we can see that all the jobs are restored after cluster is started. Does this use savepoints underhood to save and restore the jobs, or is it recommended to manually stop the jobs with savepoints before stopping the cluster, and then restore the jobs with savepoints after starting the cluster again.Thanks.

0

There are 0 answers