diff options
author | Martin Ritchie <ritchiem@apache.org> | 2007-01-12 11:19:22 +0000 |
---|---|---|
committer | Martin Ritchie <ritchiem@apache.org> | 2007-01-12 11:19:22 +0000 |
commit | 12829a2710e34bb32b7b36d7d62711a0405fe1a2 (patch) | |
tree | e5b2b30b6f6aba1a0105449740ce9e84a1c5d6ca | |
parent | b621a52c9d4efc9c80189960d0515a85687829fc (diff) | |
download | qpid-python-12829a2710e34bb32b7b36d7d62711a0405fe1a2.tar.gz |
QPID-283
Moved Resource Readme content relating to Tests to RunningPerformanceTests.txt
git-svn-id: https://svn.apache.org/repos/asf/incubator/qpid/trunk/qpid@495556 13f79535-47bb-0310-9956-ffa450edef68
-rw-r--r-- | java/resources/README.txt | 64 |
1 files changed, 0 insertions, 64 deletions
diff --git a/java/resources/README.txt b/java/resources/README.txt index 14706170bc..1d52d487fb 100644 --- a/java/resources/README.txt +++ b/java/resources/README.txt @@ -37,68 +37,4 @@ ${QPID_HOME}/bin on your PATH. The scripts in that directory include the standard ones in the distribution and a number of testing scripts. -Running Tests -------------- -The simplest test to ensure everything is working is the "service -request reply" test. This involves one client that is known as a -"service provider" and it listens on a well-known queue for -requests. Another client, known as the "service requester" creates a -private (temporary) response queue, creates a message with the private -response queue set as the "reply to" field and then publishes the -message to the well known service queue. The test allows you to time -how long it takes to send messages and receive the response back. It -also allows varying of the message size. - -You must start the service provider first: - -serviceProvidingClient.sh nop host:port - -where host:port is the host and port you are running the broker -on. - -To run the service requester: - -serviceRequestingClient.sh nop host:post <count> <bytes> - -This requests <count> messages, each of size <bytes>. After -receiving all the messages the client outputs the rate it achieved. - -A more realistic test is the "headers test", which tests the -performance of routing messages based on message headers to a -configurable number of clients (e.g. 50). A publisher sends 10000 -messages to each client and waits to receive a message from each -client when it has received all the messages. - -You run the listener processes first: - -run_many.sh 10 header "headersListener.sh -host 10.0.0.1 -port 5672" - -In this command, the first argument means start 10 processes, the -second is just a name use in the log files generated and the third -argument is the command to run. In this case it runs another shell -script but it could be anything. - -Then run the publisher process: - -headersPublisher.sh -host 10.0.0.1 -port 5672 10000 10 - -The last two arguments are: the number of messages to send to each -client, and the number of clients. - -Note that before starting the publisher you should wait about 30 -seconds to ensure all the clients are registered with the broker (you -can see this from the broker output). Otherwise the numbers will be -slightly skewed. - -A third useful test, which can easily be ported to other JMS -implementations is the "topic test". It does the same as the headers -test but using a standard topic (e.g. pub sub). - -To run the listeners: - -run_many.sh 10 topic "topicListener.sh -host 10.0.0.1 -port 5672" - -and to run the publisher: - -topicPublisher.sh -host 10.0.0.1 -port 5672 -clients 10 -messages 10000 |