loadUI Agent load problems on Linux
Hi All,
I'm trying to setup a load test but facing problems when trying to use more than 3000 concurrent VUs.
Here is what I've done (times 3 as I'm trying to use 3 agents):
Config: [ Intel E5540 with 3GB memory ]
- Installed CentOS 6.3
- Installed loadUI Agent 2.11
- Started the process /opt/SmartBear/loadUI-Agent-2.1.1/loadUI-Agent-2.1.1
- Installed loadUI on a test box
- Used an Existing SoapUI Pro Project and soapUI Pro setting XML
- Added the Agents in loadUI
- Added Ramp UP VU and soapUI Runner
When I configure more than 3000 VUs (i.e. 5000) the process becomes sluggish and the number of Discarded requests start to increase.
I've also tried with only 1 agent and the result is consistent i.e 1 agent will work with 1000 VUs
Ulimit -n and ulimit -a are both over 10000 (both client and Agent)
Observation on the target server:
I can see that the target machine is not receiving the desired number of requests.
Observation on the client (loadUI):
- All 3 machines have a high load
- From shell using top CPU is 100% on the Java process
* Can someone please confirm that my setup is correct?
* What could be the cause of this problem? (tried setting ulimit higher but didn't seem to help)
I'm trying to setup a load test but facing problems when trying to use more than 3000 concurrent VUs.
Here is what I've done (times 3 as I'm trying to use 3 agents):
Config: [ Intel E5540 with 3GB memory ]
- Installed CentOS 6.3
- Installed loadUI Agent 2.11
- Started the process /opt/SmartBear/loadUI-Agent-2.1.1/loadUI-Agent-2.1.1
- Installed loadUI on a test box
- Used an Existing SoapUI Pro Project and soapUI Pro setting XML
- Added the Agents in loadUI
- Added Ramp UP VU and soapUI Runner
When I configure more than 3000 VUs (i.e. 5000) the process becomes sluggish and the number of Discarded requests start to increase.
I've also tried with only 1 agent and the result is consistent i.e 1 agent will work with 1000 VUs
Ulimit -n and ulimit -a are both over 10000 (both client and Agent)
Observation on the target server:
I can see that the target machine is not receiving the desired number of requests.
Observation on the client (loadUI):
- All 3 machines have a high load
- From shell using top CPU is 100% on the Java process
* Can someone please confirm that my setup is correct?
* What could be the cause of this problem? (tried setting ulimit higher but didn't seem to help)
