Druid Coordinator isn't creating kill tasks

[Henrique Martins]

Good afternoon,

Recently I updated a project to work with those configurations:

druid_coordinator_kill_on=true
druid_coordinator_kill_period=PT1800S
druid_coordinator_kill_ignoreDurationToRetain=true
druid_coordinator_kill_bufferPeriod=PT0S

It's running inside a docker. I do't understand why it's not working. The retention rules mark the segments as unused but Coordinator never creates a kill task.

Br,

Henrique Martins.

[Henrique Martins]

I solved the issue. We are working on a PoC. Our project hade Middle Manager work capacity set to 2. That means each task uses 50% of the worker capacity. Each task run in a Java VM.

Our killTaskSlotRatio was set to 0.1(default value). That means for our Coordinator to create a kill task it had only 10% of worker capacity. But for a task to run it needs at least 50%.

For that reason, even when we weren't running tasks it wasn't creating kill tasks. After changing that value we could see kill tasks being generated.

The killTaskSlotRatio wasn't ok.

Problem is solved.