[
https://issues.apache.org/jira/browse/CASSANDRA-13843?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16154982#comment-16154982
]
Simon Fontana Oscarsson commented on CASSANDRA-13843:
-----------------------------------------------------
I would rather solve this issue with CASSANDRA-13006. So instead of letting
Cassandra creating it's own heapdumps leave that job to the JVM.
> Debian init shadows CASSANDRA_HEAPDUMP_DIR
> ------------------------------------------
>
> Key: CASSANDRA-13843
> URL: https://issues.apache.org/jira/browse/CASSANDRA-13843
> Project: Cassandra
> Issue Type: Bug
> Components: Observability
> Reporter: Andrew Jorgensen
> Priority: Minor
> Labels: newbie, patch
> Fix For: 3.11.x, 4.x
>
> Attachments: 0001-Remove-debian-init-setting-heap-dump-file.patch
>
>
> The debian init script sets the heap dump file directly using the cassandra
> users home directory and the -H flag to the cassandra
> process[1|https://github.com/apache/cassandra/blob/8b3a60b9a7dbefeecc06bace617279612ec7092d/debian/init#L76].
> The cassandra heap dump location can also be set in the cassandra-env.sh
> file using CASSANDRA_HEAPDUMP_DIR. Unfortunately the debian init heap dump
> location is based off the home directory of the cassandra user and cannot
> easily be changed. Also unfortunately if you do `ps aux | grep casandra` you
> can clearly see that the -H flag takes precedent over the value found in
> cassandra-env.sh. This makes it difficult to change the heap dump location
> for cassandra and is non-intuitive when the value is set in cassandra-env.sh
> why the heap dump does not actually end up in the correct place.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
