[
https://issues.apache.org/jira/browse/FALCON-1623?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15204409#comment-15204409
]
Balu Vellanki commented on FALCON-1623:
---------------------------------------
[~peeyushb] : What happens to running instances of Feed/Process during cluster
update is explained in the cluster update doc. Essentially it is same as
updating a feed/process ... existing coord/bundle are updated to change endtime
to now. New coord/bundle will be created with start time as now. Safemode in
Falcon has no implication on what happens to jobs in the workflow engine.
You bring up good points about LateRerunHandler and RetryHandler. Falcon will
not schedule any jobs in workflow engine during safemode. This should include
retry/rerun jobs. I will update doc accordingly. MetadataMappingService,
notifications etc will continue in safemode as these are not updating wfe.
> Implement Safe Mode in Falcon
> -----------------------------
>
> Key: FALCON-1623
> URL: https://issues.apache.org/jira/browse/FALCON-1623
> Project: Falcon
> Issue Type: Bug
> Reporter: sandeep samudrala
> Assignee: Balu Vellanki
> Attachments: FALCON-1623.patch, FalconServerSafeMode..docx, Safe Mode
> in Falcon.pdf
>
>
> Safe mode needs to be implemented in falcon to let falcon go into safe mode
> during outages and also help cluster update .
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
