[
https://issues.apache.org/jira/browse/SPARK-8734?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14933403#comment-14933403
]
Ondřej Smola commented on SPARK-8734:
-------------------------------------
The goal should be to try provide as much value as we can given this limited
format, as i said i think from what other frameworks provide in their RESP API
we need to handle in this stage at least ports,volumes, network type, env
variables, labels - without limitation. So proposed model is
spark.mesos.executor.docker.network host
spark.mesos.executor.docker.portmaps 1234:1245:udp, 2546:1255:tpc
spark.mesos.executor.docker.volumes /usr/:/host/usr:ro, /opt:/host/opt
spark.mesos.executor.docker.env.FOO BAR
spark.mesos.executor.docker.label.STAGE PRODUCTION
Support for custom docker parameters
spark.mesos.executor.docker.parameter.memory-swappiness 0
Limitation - only one value for key is allowed
spark.mesos.executor.docker.parameters.attach stdin,stdout
Limitation - comma can be used only as a separator
Pros
consistent with other configuration sections
we can easily move custom parameters to own section in the future
no escaping problems
Cons
Not powerful enough to express some docker parameter combinations (multiple
values containing commas)
> Expose all Mesos DockerInfo options to Spark
> --------------------------------------------
>
> Key: SPARK-8734
> URL: https://issues.apache.org/jira/browse/SPARK-8734
> Project: Spark
> Issue Type: Improvement
> Components: Mesos
> Reporter: Chris Heller
> Priority: Minor
> Attachments: network.diff
>
>
> SPARK-2691 only exposed a few options from the DockerInfo message. It would
> be reasonable to expose them all, especially given one can now specify
> arbitrary parameters to docker.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
