[
https://issues.apache.org/jira/browse/CASSANDRA-5067?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13570748#comment-13570748
]
Brandon Williams commented on CASSANDRA-5067:
---------------------------------------------
If people want to configure things this way, I'd be ok with that, but it's not
simple to do. For every property we allow overriding, we're going to have to
explicitly define it. Every time we add a new property, we'll have to think
about whether it needs an override as well, and probably forget to add it and
have to do it later. There's no way for us to provide a generic namespace to
override system properties, since even if we scanned through them all and
filtered by prefix we wouldn't know how to cast them to the different values
needed, not to mention in some cases like data_file_directories we'd have to
use CSV or something to represent an array. So there is a lot of technical
debt to be incurred by doing this, even though it seems simple on the surface.
The only way to really do this would be to have another yaml file which could
override the first, and merge them at startup. But then we end up with 2x the
files and you'd still ultimately be writing the 'override file'
programmatically, so I don't see much point in going that route, either.
> Allow basic yaml properties to be specified via JVM parameters
> --------------------------------------------------------------
>
> Key: CASSANDRA-5067
> URL: https://issues.apache.org/jira/browse/CASSANDRA-5067
> Project: Cassandra
> Issue Type: Improvement
> Reporter: Ahmed Bashir
> Priority: Trivial
> Labels: ponies
>
> In the case of large clusters, each node operates with a cassandra.yaml that
> is nearly identical to every other node with the exception of some basic
> properties:
> listen_address
> rpc_address
> initial_token
> data_file_directories
> etc
> If it was possible to specific these as JVM parameters, we wouldn't have to
> prepare a customized cassandra.yaml for each node but rather share a single
> cassandra.yaml file. This is increasingly important when the number of nodes
> being managed gets very large.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
