It sounds to me like the client isn't picking up the gemfire.properties file when run under JMeter, so it doesn't know the address of the locator.
-- Mike Stolz Principal Engineer, GemFire Product Manager Mobile: 631-835-4771 On Tue, Nov 10, 2015 at 5:35 PM, Vincent Ford <[email protected]> wrote: > As mentioned by Barry, you should capture thread dumps to see where things > are stuck. Also increasing the logging level to fine may also give hints as > to the cause of the issue... Are you making any other settings to the > client cache or locator? > > *Vince Ford* > GemFire Sustenance Engineering > Beaverton, OR USA > 503-533-3726 (office) > http://www.pivotal.io > Open Source Project Geode https://geode.incubator.apache.org/ > <https://network.pivotal.io/products/project-geode> > > On Tue, Nov 10, 2015 at 9:27 AM, Barry Oglesby <[email protected]> > wrote: > >> JMeter has been used to run perf tests with GemFire in the past. I >> haven't seen it used with Geode yet, but it should work ok. I guess I would >> start by getting a thread dump in the client JVM to see where in the cache >> creation process it is stuck. >> >> Barry Oglesby >> GemFire Advanced Customer Engineering (ACE) >> For immediate support please contact Pivotal Support at >> http://support.pivotal.io/ >> >> >> On Tue, Nov 10, 2015 at 7:18 AM, Matt Ross <[email protected]> wrote: >> >>> Hi All, >>> >>> I am currently trying to do Unit Tests on my Gemfire Cluster using the >>> load testing tool Apache JMeter. I have a spring boot application that >>> creates a client connection to a gemfire cluster, runs a simple query and >>> then disconnects. This app runs just fine when I run the JUnit tests in >>> eclipse, but when I run it through JMeter the client creation never occurs. >>> >>> >>> For some reason when the spring boot app is making the connection from >>> behind JMeter, I encounter a problem when I make the following call: >>> >>> * cache = new ClientCacheFactory()* >>> >>> The client just hangs. My tests don't throw a runtime exception, but >>> they also don't move past the step of creating the cache object. I checked >>> the server logs and saw nothing, and when I checked the locator logs I had >>> this exception being repeatedly thrown. >>> >>> *[info 2015/11/09 09:56:26.491 EST locator1 <UDP ucast receiver> >>> tid=0x26] failed unmarshalling message from >>> matt(locator1:3658:locator)<v0>:41019 java.lang.NullPointerException* >>> >>> So I shutdown my cluster and set mcast-port=0 on my locator and servers. >>> Then when I ran the JMeter Unit tests again, the null pointer exception >>> wasn't being thrown anymore, but the client-cache hangs and still doesn't >>> make the connection. In order to see what was happening at that cache >>> initialization stage, I gave it a custom path for where to write the >>> client-cache logs to, but the client cache doesn't even write any logs. >>> >>> Has anyone ever worked with Gemfire and JMeter before and encountered >>> this type of issue? Any input would be appreciated. Thanks >>> -- >>> Matthew Ross | Pivotal Data Engineering | 516-941-7535 >>> >> >> >
