Hi,

Currently system like Storm only support kill-and-resubmit kind of job
swapping for a live system, which is not good as down time can be up to 10
minutes in our experience.

Is this issue better handled in Spark Streaming, such as with the help of
mesos?

Reply via email to