Chronos not respecting role parameter

[bdonnahue]

I have a cluster with a single master node and a single agent node of type slave_public. I cannot get chronos to deploy because it does not respect the settings in the json file for the app. Specifically "acceptedResourceRoles":["slave_public"] as mentioned here: https://docs.mesosphere.com/1.7/usage/tutorials/public-app/. Despite this setting my logs still show the following: Oct 21 18:58:25 MesosMaster marathon: [2018-10-21 18:58:25,280] #033[34mINFO #033[0;39m Offer [84b9fe28-fb90-4b2c-894f-99fb10b9ceb2-O172]. Considering resources with roles {*} without resident reservation labels. Not all basic resources satisfied: cpus NOT SATISFIED (0.5 > 0.0), mem NOT SATISFIED (1024.0 > 0.0) (mesosphere.mesos.ResourceMatcher$:marathon-akka.actor.default-dispatcher-56) If the setting was being respected I believe the logs would show a different message " ... with roles {"slave_master"} ... "

If this behavior is there by design, I think the documentation should be updated to reflect the fact that this app only runs on a private cluster.

[gkleiman]

@bdonnahue those look like Marathon logs, not Chronos logs.