The load is not generated from mulitple client instances as expected in distributed environment

Hello,

I am using Tsung tool to perform XMPP load testing in distributed environment and I am not getting the load from each client properly.

I have three client machines and I increased the open files, max user process and the file max values already. Also I can able to run like 30k concurrent users from single machine.

Below is the client configuration which I given in my script,

    <client host="localhost" maxusers="64000" use_controller_vm="true">
    <ip value="10.xxx.x.xx"/>
    </client>

    <client host="tsung1" maxusers="64000" use_controller_vm="true">
    <ip value="10.xxx.x.xx"/>
    </client>

    <client host="tsung2" maxusers="64000" use_controller_vm="true">
    <ip value="10.xxx.x.xx"/>
    </client>

The load progressed as,

 <load>
<arrivalphase phase="1" duration="10" unit="second">
    <users maxnumber="60000" arrivalrate="100" unit="second"/>
</arrivalphase>
</load>

The jabber details,

I get the username and password from a CSV file to login to the application using .

   <setdynvars sourcetype="file" fileid="users" delimiter="," order="iter">
       <var name="username"/>
       <var name="password"/>
     </setdynvars>

Environment Details:

OS - Ubuntu 18.04
CPU - 8 vCPU
RAM - 16 GB

Following is the tsung controller log,

=INFO REPORT==== 23-Sep-2020::14:40:53 === ts_mon:(5:<0.102.0>) Starting tsung clients on hosts: [tsung2, tsung1, localhost]

=INFO REPORT==== 23-Sep-2020::14:40:53 === ts_config_server:(5:<0.101.0>) Start a launcher on the controller beam localhost

=INFO REPORT==== 23-Sep-2020::14:40:53 === alarm_handler: {set,{{disk_almost_full,"/snap/core18/1885"},[]}}

=INFO REPORT==== 23-Sep-2020::14:40:53 === alarm_handler: {set,{{disk_almost_full,"/snap/google-cloud-sdk/151"},[]}}

=INFO REPORT==== 23-Sep-2020::14:40:53 === alarm_handler: {set,{{disk_almost_full,"/snap/snapd/8790"},[]}}

=INFO REPORT==== 23-Sep-2020::14:40:53 === alarm_handler: {set,{{disk_almost_full,"/snap/google-cloud-sdk/150"},[]}}

=INFO REPORT==== 23-Sep-2020::14:40:53 === alarm_handler: {set,{{disk_almost_full,"/snap/snapd/9279"},[]}}

=INFO REPORT==== 23-Sep-2020::14:40:53 === ts_launcher_static:(5:<0.156.0>) Launch msg receive ("localhost")

=INFO REPORT==== 23-Sep-2020::14:40:53 === ts_launcher:(5:<0.157.0>) Launch msg receive ("localhost")

=INFO REPORT==== 23-Sep-2020::14:40:53 === ts_user_server:(5:<0.116.0>) Reset offline and online lists (maxid=60000)

=INFO REPORT==== 23-Sep-2020::14:40:53 === ts_config_server:(5:<0.158.0>) Remote beam started on node tsung1@tsung2

=INFO REPORT==== 23-Sep-2020::14:40:53 === ts_config_server:(5:<0.159.0>) Remote beam started on node tsung2@tsung1

=INFO REPORT==== 23-Sep-2020::14:40:53 === ts_config_server:(5:<0.101.0>) All remote beams started, syncing

=INFO REPORT==== 23-Sep-2020::14:40:54 === ts_config_server:(5:<0.101.0>) Undefined ports_range config

=INFO REPORT==== 23-Sep-2020::14:40:54 === ts_config_server:(5:<0.101.0>) New arrival phase 1 for client "localhost" (last ? false): will start 333 users

=INFO REPORT==== 23-Sep-2020::14:40:54 === ts_launcher:(5:<0.157.0>) Expected duration of first phase: 10.0 sec (333 users)

=INFO REPORT==== 23-Sep-2020::14:40:54 === ts_config_server:(5:<0.101.0>) New arrival phase 1 for client "tsung2" (last ? false): will start 333 users

=INFO REPORT==== 23-Sep-2020::14:40:54 === ts_launcher:(5:<0.157.0>) Activate launcher (333 users) in 1000 msec

=INFO REPORT==== 23-Sep-2020::14:40:54 === ts_config_server:(5:<0.101.0>) New arrival phase 1 for client "tsung1" (last ? true): will start 333 users

=INFO REPORT==== 23-Sep-2020::14:41:05 === ts_launcher:(5:<0.157.0>) This was the last phase, wait for connected users to finish their session

Can anyone please look into this issue and help on the same.

Note: Last year December, I achieved the distribution load testing with 100k users in CentOS.

processone / tsung

The load is not generated from mulitple client instances as expected in distributed environment #381