Running Multiple Indexing tasks failing

When I have more than one ingestion spec running, the tasks fail in an alternating fashion. It seems that they’re stepping on each other’s toes, but I can’t quite determine what it is (nothing shows up in the error log). The pattern seems to be indexing task 1 starts up, then indexing task 2 starts up. As soon as task 2 starts up, it seems to kill task 1. I can’t see why it would be doing this.

I’m running druid inside docker images on kubernetes, so I imagine it has something to do with the setup there, but I can’t for the life of me figure it out.

Nothing inside the task logs seems weird (this is the end of the task log):

2017-11-01T16:47:42,815 INFO [main]$AnnotationBasedHandler - Invoking start method[public void io.druid.server.listener.announcer.ListenerResourceAnnouncer.start()] on object[io.druid.query.lookup.LookupResourceListenerAnnouncer@5e5ddfbc].

2017-11-01T16:47:42,819 INFO [main] io.druid.server.listener.announcer.ListenerResourceAnnouncer - Announcing start time on [/druid/listeners/lookups/__default/druid-middlemanager-service-0.druid-middlemanager-service.staging.svc.cluster.local:8100]


Any help would be much appreciated!

Maybe you don’t have enough memory to launch more than one task?