# Esrally hangs dead **URL:** \ **Category:** Elasticsearch\ **Tags:** rally\ **Created:** [July 21, 2019, 2:05am UTC](https://discuss.elastic.co/t/esrally-hangs-dead/191498 "2019-07-21T02:05:20Z")\ **Posts on this page:** 7\ **Page:** 1
**Author:** ![baoshengzhu](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/baoshengzhu/32/50587_2.png) [@baoshengzhu](https://discuss.elastic.co/u/baoshengzhu)\ **Post date:** [July 21, 2019, 2:05am UTC](https://discuss.elastic.co/t/esrally-hangs-dead/191498/1 "2019-07-21T02:05:20Z")
when I set bulk\_indexing\_clients equal to 500, esrally hangs dead, I need to set higher value of clients, how to solve? ---
**Author:** ![Christian\_Dahlqvist](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/christian_dahlqvist/32/4617_2.png) [@Christian\_Dahlqvist](https://discuss.elastic.co/u/Christian_Dahlqvist)\ **Post date:** [July 21, 2019, 8:07am UTC](https://discuss.elastic.co/t/esrally-hangs-dead/191498/2 "2019-07-21T08:07:22Z")
Why do you need so many clients? What is the size of the cluster you are benchmarking? What is the specification of the host it is running on? Which track are you using? Esrally is implemented in Python and uses a separate process per client. If you therefore want to use 500 clients you most likely need to run it in distributed mode. ---
**Author:** ![baoshengzhu](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/baoshengzhu/32/50587_2.png) [@baoshengzhu](https://discuss.elastic.co/u/baoshengzhu)\ **Post date:** [July 21, 2019, 9:17am UTC](https://discuss.elastic.co/t/esrally-hangs-dead/191498/3 "2019-07-21T09:17:15Z")
> [@Christian\_Dahlqvist](#): > > distributed mode. Our elasticsearch cluster is a server with 50 SSD disks and currently uses esrally distributed pressure measurement，But I set bulk\_indexing\_clients equal to 500, esrally hangs dead, When I turned it down to 280 it worked perfectly. ---
**Author:** ![Christian\_Dahlqvist](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/christian_dahlqvist/32/4617_2.png) [@Christian\_Dahlqvist](https://discuss.elastic.co/u/Christian_Dahlqvist)\ **Post date:** [July 21, 2019, 9:28pm UTC](https://discuss.elastic.co/t/esrally-hangs-dead/191498/4 "2019-07-21T21:28:24Z")
What is the specification of the server you are benchmarking? What is the purpose of the benchmark? Which track are you using? ---
**Author:** ![baoshengzhu](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/baoshengzhu/32/50587_2.png) [@baoshengzhu](https://discuss.elastic.co/u/baoshengzhu)\ **Post date:** [July 22, 2019, 4:54pm UTC](https://discuss.elastic.co/t/esrally-hangs-dead/191498/5 "2019-07-22T16:54:58Z")
We are using the http\_log data set for benchmarking, with a target throughput of 4 million documents per second.Is the bulk\_indexing\_clients setting limited? ---
**Author:** ![Christian\_Dahlqvist](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/christian_dahlqvist/32/4617_2.png) [@Christian\_Dahlqvist](https://discuss.elastic.co/u/Christian_Dahlqvist)\ **Post date:** [July 22, 2019, 5:28pm UTC](https://discuss.elastic.co/t/esrally-hangs-dead/191498/6 "2019-07-22T17:28:55Z")
I would recommend you start much lower and gradually increase the number of clients until you see no further throughput improvement. I would expect to be able to saturate a cluster using far rewear clients. ---
**Author:** ![system](https://us1.discourse-cdn.com/elastic/original/3X/1/a/1ac57faf039f6b580b3f104ef42a2a89e41014de.png) [@system](https://discuss.elastic.co/u/system)\ **Post date:** [August 19, 2019, 5:28pm UTC](https://discuss.elastic.co/t/esrally-hangs-dead/191498/7 "2019-08-19T17:28:56Z")
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.