Bulk Index task hangs

daniel.b.sario · August 10, 2018, 4:14am

I did a bulk index task using java api. After a few successful bulk index requests, suddenly my indexer hangs. I checked the the status of index tasks using the task api of elasticsearch and I found that there are tasks still running and was going for 30+ minutes and its status is 'waiting_on_primary'.

What should I do to fix this?, Do I need to change something from the elasticsearch.yml?. Please HELP!

dadoonet · August 10, 2018, 5:54am

Anything in logs?

Christian_Dahlqvist · August 10, 2018, 6:19am

What is the size of your documents? How large are your bulk requests? How many parallel indexing threads/processes do you have? What is the specification of the cluster?

daniel.b.sario · August 10, 2018, 6:53am

@Christian_Dahlqvist I have the below configuration for BulkProcessor

Bulk Size = 5MB
Flush Interval = 30 seconds
Concurrent Request = 2

I only have one node, 8 shards. the ES version im using is 6.2.3

daniel.b.sario · August 10, 2018, 6:58am

@dadoonet There were nothing on logs except for those documents that failed in indexing. I cant cancel the task as is tagged as not cancel-able.

system · September 7, 2018, 6:58am

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.