Druid Ingestion and Segment Load issues

[kashmira aher]

Hi Everyone, 

I have 2 node setup in druid, I have being using it for 3 months now and it is giving an  error for ingestion jobs as "heap out of memory" how should I handle this or is there a need to increase the VM's

And if the ingestion get success it takes 2-3 days to reflect the data on Superset

The compaction are running continuously for every datasource but now it is failing too.

The segements get unavailable and datasources show an error as below:

"Unknown exception// by zero/java.lang.ArithmticException"

Please suggest what can be done?

Thank You

Regards

Kashmira Aher.

[Peter Marshall]

Hey Kashmira :)

The arithmetic exception could be this:

https://github.com/apache/druid/issues/10507

Heap sizing information is in the tuning guide - you need to check which process is out of Heap space - for example, if you now have a lot of segments, then the broker and / or coordinator might be unable to fit that all into memory now.

https://druid.apache.org/docs/0.20.2/operations/basic-cluster-tuning.html#process-specific-guidelines

As for the delay in appearing to superset - then you should check the load queues - if you cannot get to the console's page (because of that error) there are a number of APIs you can use:

https://druid.apache.org/docs/0.20.2/operations/api-reference.html#segment-loading