I am trying do distributed testing on linux server using apache-jmeter 2.9
The default port (1099) is already used(by jboss)
I changed the port as 1097
I start jmeter-server on one machine for now and start test on single machine.
jmeter-server seems to start succesfuly
but when evern i trying to exceute script is shows following error.
[jboss#StagingSvr2 bin]$ ./jmeter -n -t CBL_Load/CBL_Admin_Load.jmx -l
.jtl -R 172.16.0.2
Creating summariser <summary>
Created the tree successfully using CBL_Load/CBL_Admin_Load.jmx
Configuring remote engine for 172.16.0.2
Failure connecting to remote host: 172.16.0.2
java.rmi.ConnectIOException: non-JRMP server at remote endpoint
Failed to configure 172.16.0.2
[![enter image description here][1]][1]No remote engines were started.
I have gone through google but not able to find exact solution that where I am doing blundder!
Make sure nothing is listening at the port 1097 using netstat or nc or telnet. Looking into non-JRMP server at remote endpoint something is present there which is not JMeter RMI endpoint. Try locating a free port using aforementioned tools and bind JMeter slave to it
With regards to bind JMeter slave I would recommend amending your startup command to something like:
./jmeter-server -Dserver_port=xxxx
where xxxx is a free port on your Linux system
Amend your Master startup command to include the port as well like:
./jmeter -R 172.16.0.2:xxxx -n -t CBL_Load/CBL_Admin_Load.jmx -l result.jtl
More information:
JMeter Remote Testing: Using a different port
How to Perform Distributed Testing in JMeter
Related
Jmeter Distributed testing - unable to connect the remote (slave systems)
I am getting the below error
"non-JRMP server at remote endpoint"
It looks like an incorrect setup of RMI ports, JMeter master tries to connect to the slave machine using the hostname or IP address and port which is specified in the remote_hosts property or passed via -R command line argument.
The error means that something else, not JMeter's RMI registry is running on machine 192.168.0.103 on port 1099.
So you need to connect to your slave machine and check what's listening on port 1099 using netstat or equivalent and either terminate that application or amend JMeter configuration to use different port.
More information: How to Perform Distributed Testing in JMeter
In centos environment I setup a rmi_keystore.jks in master device and copy it to the bin of Jmeter worker machine too.
In my master device's jmeter.properties file I only did below changes.
remote_hosts=10.54.225.200
server.rmi.localport=4000
However, In worker device, when I try to start the server.
by giving master's hostname is 10.54.225.190
cd apache-jmeter-5.4.1/bin
./jmeter-server -Djava.rmi.server.hostname=10.54.225.190
below error occurs. The port 4000 is not in used though.
If you need to use different SERVER_PORT instead of default 1099 I believe you need to amend your setup to something like:
Slave: jmeter-server -Jserver_port=4000
Master: jmeter -R 10.54.225.200:4000 -n -t test.jmx
See Using a different port user manual entry for more details if needed.
If you want to customize further refer the following materials:
Apache JMeter Distributed Testing Step-by-step
Remote hosts and RMI configuration
JMeter Distributed Testing with Docker
I'am trying to use Hue as a file browser for HDFS. So for that I have clone the hue repository and build the app with the following commands given in README.md of the hue repository.
git clone https://github.com/cloudera/hue.git
cd hue
make apps
build/env/bin/hue runserver
Hue UI is accessible in local machine using default port using the url http://localhost:8000 and everything works fine. But when I use my machine ip address http://x.x.x.x:8000 and try to access the Hue UI it keeps on processing and waiting.
Other observations -:
I can ping from remote machine to the host machine.
There is no firewall blocking the ports. (checked with nmap port scanner)
Machines are in same network.
I can access other ports for Hadoop NameNodes UI and DataNodes.
Changing the http_host in hue.ini doesn't affect the result
The ideal setup for Hue is configuring a reverse proxy (Nginx or Apache HTTP, for example)
However, you should refer to the Configuration documentation to externally run the server outside of 127.0.0.1
[desktop]
# Webserver listens on this address and port
http_host=0.0.0.0
http_port=8888
I was able to find a solution to the issue.. First hue run on a CherryPy web server so starting server by command build/env/bin/hue runserver will start the development server where hue.ini configuration is neglected.
So the correct command to start the production server after setting up correct configuration in hue.ini file is build/env/bin/hue runcpserver. Then I was able to access it using remote host without any problem. You also can use supervisor to start the production server. More information about that can be found here
Why am I getting the following error in PerfMon jmeter plugin?
"ERROR: java.io.IOException: Agent is unreachable via TCP"
how to configure perfmon plugin in Jmeter 4.0 version
Make sure to install Server Agent to the server you want to monitor and launch it
Make sure that default port of 4444 is open for TCP and UDP traffic in your operating system firewall.
In order to test connectivity you can send test message to the Server Agent using Telnet client software (you can use Putty if you don't have any Telnet client installed). If everything is fine you should see Yep response. JMeter should be able to connect to the Server Agent and collect the metrics of your choice:
More information: How to Monitor Your Server Health & Performance During a JMeter Load Test
I have installed Cloud 9 IDE on my server, but I can't access to it remotely.
info - socket.io started
Project root is: .
Trying to start your browser in: http://127.0.0.1:3000
Does Cloud 9 work only on local?
Yes, by default the Cloud9 webserver listens only to the localhost interface, to specify another address, just add the following argument to the command line: -l 0.0.0.0, e.g.
./cloud9.sh -l 0.0.0.0 -w ~/workspace