[
https://issues.apache.org/jira/browse/NIFI-1238?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15035120#comment-15035120
]
Mark Payne commented on NIFI-1238:
----------------------------------
An easy way to replicate this issue is to use the ForceOOME.xml template
attached, and have the GetFile processor ingest hello.txt. This will cause the
flow to generate FlowFiles with millions of lines each and then route them to
SplitText. This will cause massive amounts of garbage collection (and generally
OOMEs). (There is already another ticket to address this problem).
When trying to call "bin/nifi.sh stop" while the SplitText processor is
running, NiFi is automatically restarted.
> If "nifi.sh stop" is run and the connection times out, nifi is restarted
> instead of stopped
> -------------------------------------------------------------------------------------------
>
> Key: NIFI-1238
> URL: https://issues.apache.org/jira/browse/NIFI-1238
> Project: Apache NiFi
> Issue Type: Bug
> Components: Core Framework
> Affects Versions: 0.4.0
> Reporter: Mark Payne
> Assignee: Mark Payne
> Fix For: 0.4.0
>
> Attachments: ForceOOME.xml, hello.txt
>
>
> When we run the stop script, if the bootstrap module is unable to communicate
> with NiFi, it will issue a "kill -9" command (for *nix/osx). However, the
> already-running bootstrap process will then detect that NiFi died and restart
> it.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
