Mikhail Pochatkin created IGNITE-20850:
------------------------------------------
Summary: Worker node shutdown failover
Key: IGNITE-20850
URL: https://issues.apache.org/jira/browse/IGNITE-20850
Project: Ignite
Issue Type: Improvement
Components: compute
Reporter: Mikhail Pochatkin
In this case, the job execution stops and you need to restart it. In this case,
the coordinator will see that the worker node has turned off and all the tasks
that this coordinator sent for execution must be redistributed to other nodes.
In this context, it does not matter what state the tasks were in, in the queue
or in the process of execution, we do not offer a safepoint mechanism, and all
states of jobs can be written to the cache, so when a task is launched on
another worker node, it will be able to read from the cache the state that the
job I wrote it down last time.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
