Environment for creating a container image running cph performance tests for Persistent and Non Persistent MQ messaging.
This repository contains a set of files to help create a container image containing the CPH executable and a set of scripts to run an inital set of performance tests.
The Java version of this testharness using JMS interfaces is available here: https://github.com/ibm-messaging/jmstestp
You will need to seperately download the MQ Client (for which license agreement is required) and copy the following files into the root directory before building your container image:
- /lap/
- mqlicense.sh
- ibmmq-client_9.x.x.x_amd64.deb
- ibmmq-gskit_9.x.x.x_amd64.deb
- ibmmq-runtime_9.x.x.x_amd64.deb
The MQ V9 client can be obtained from: http://www-01.ibm.com/support/docview.wss?uid=swg24042176
The MQ V9.1 client can be obtained from: http://www-01.ibm.com/support/docview.wss?uid=swg24044791
The MQ V9.2 client can be obtained from: https://ibm.biz/mq92clients
The MQ V9.3 client can be obtained from: https://ibm.biz/mq93clients
The MQ V9.4 client can be obtained from: https://ibm.biz/mq94clients
For the above downloads, filter the fixpack selection platform to Linux 64 bit, x86_64, and select the IBM-MQC-UbuntuLinuxX64 download.
The Dockerfile will need to be edited to refer to the client version you have downloaded, then perform a container build as normal:
docker build --tag cphtestp .
then run in network host mode to connect and run tests against a local QM:
docker run -it --detach --net="host" cphtestp
The default configuration looks for a QM located on the localhost called PERF0 with a listener configured on port 1420. The clients will send and receive persistent messages. You can override a number of options by setting environment variables on the container run command.
docker run -it --detach --net="host" --env MQ_QMGR_NAME=PERF1 --env MQ_QMGR_HOSTNAME=10.0.0.1 --env MQ_QMGR_PORT=1414 --env MQ_QMGR_CHANNEL=SYSTEM.DEF.SVRCONN --env MQ_QMGR_QREQUEST_PREFIX=REQUEST --env MQ_QMGR_QREPLY_PREFIX=REPLY cphtestp
In addition to the hostname, port and and QM name, the default channel can be overidden using the MQ_QMGR_CHANNEL envvar and the queue prefixes used for the testing can be set using MQ_QMGR_QREQUEST_PREFIX and MQ_QMGR_QREPLY_PREFIX.
You can also run this performance harness inside Red Hat OpenShift using the OpenShift instructions.
In the latest release further configuration options have been added. The table below provides the full set:
| Envvar | Description | Default if not set |
|---|---|---|
| MQ_QMGR_NAME | Queue Manager Name | PERF0 |
| MQ_QMGR_HOSTNAME | Hostname where QM is running | localhost |
| MQ_QMGR_PORT | Port where QM listener is running | 1420 |
| MQ_QMGR_CHANNEL | Channel name to use to connect to QM | SYSTEM.DEF.SVRCONN |
| MQ_FIXED_CLIENTS | Specify a fixed number of requester clients to use | 1 2 4 8 ... <responder threads> |
| MQ_RESPONDER_THREADS | Number of responder threads to run | 200 |
| MQ_MSGSIZE | Message sizes to test | 2048:20480:204800 |
| MQ_QMGR_QREQUEST_PREFIX | Prefix of request queues to use. | REQUEST |
| MQ_QMGR_QREPLY_PREFIX | Prefix of reply queues to use. | REPLY |
| MQ_NON_PERSISTENT | QOS to be used by connecting clients | 0 (Persistent) |
| MQ_USERID | Userid to use when authenticating | |
| MQ_PASSWORD | Password to use when authenticating | |
| MQ_CPH_EXTRA | Additional string field to propogate to cph client | |
| MQ_RESULTS | Log results to stdout at end of tests | TRUE |
| MQ_RESULTS_CSV | Log results to csv file and send to stdout at end | FALSE |
| MQ_TLS_CIPHER | TLS CipherSpec to use | |
| MQ_TLS_CERTLABEL | TLS Certificate Label to use for the client | |
| MQ_TLS_SNI_HOSTNAME | Use TLS SNI field to indicate target host | FALSE |
| MQ_ERRORS | Log MQ error log at end of test | FALSE |
| MQ_RUNLENGTH | Length of each test iteration (seconds) | 90 |
| MQ_AUTORECONNECT | MQ Auto reconnection option | |
| MQ_COMPRESS | Enable client compression algorithms | FALSE |
The container will run a number of tests using different numbers of requester threads(incrementing in powers of 2) with messages of 2K, 20K and 200K. The scenario is a Request/Responder scenario as featured in the latest xLinux and Appliance performance reports available here: https://ibm-messaging.github.io/mqperf/
The final iteration will be with requester threads equal to the responder threads, thus you can control the scope of the testing by just setting the MQ_RESPONDER_THREADS envvar.
To ensure the timestamps created when testing match those of your hostmachines, you can edit /usr/share/containers/containers.conf and set tz=local to ensure that the container runs within the same timezone as that of the hostmachine.
When the testing is complete the final results will be posted to the container logs and can be viewed in the normal way:
docker logs <containerID>
Here are some snippets of the human readable MQ_RESULTS:
CPH Test Results
Tue Oct 16 15:48:22 UTC 2018
2K
threads=1
avgRate=1150.56
CPU=1.35
Read=0.00
Write=0.00
Recv=0.06
Send=0.05
QM_CPU=2.70
and also of the MQ_RESULTS_CSV output:
# Tue Oct 16 15:47:52 UTC 2018
# Persistence, Msg Size, Threads, Rate (RT/s), Client CPU, IO Read (MB/s), IO Write (MB/s), Net Recv (Gb/s), Net Send (Gb/s), QM CPU
1,2048,1,1150.56,1.35,0.00,0.00,0.06,0.05,2.70
You can also obtain the available results by:
docker cp <containerID>:/home/mqperf/cph/results .
The output from the running responder and requester processes can be viewed by:
docker cp <containerID>:/home/mqperf/cph/output .
An interactive session with the running container can be access by:
docker -ti <containerID> /bin/bash
Logging to a CSV output file is also supported by setting MQ_RESULTS_CSV to true, and the human readable output can be disabled by setting MQ_RESULTS=FALSE. The CSV output contains the date and time of the test, and field description header (both preceeded by #) before the CSV data.
Support for TLS has now been added with the MQ_TLS_CIPHER environment variable, from which a new CCDT will be created for the MQ client to use. Set the CipherSpec to match/work with the CipherSpec defined at the QM. You will need to place a CMS keystore named key.kdb(and its stash file) containing your QM public certificate in the /ssl directory.
You can also specify the Certificate Label that the cph client will use, so that mutual authentication takes place. If you leave blank, then just QM authentication will be used; in which case make sure your server channel definition specifies SSLCAUTH(OPTIONAL) and not SSLCAUTH(REQUIRED).
The use of the MQ_TLS_SNI_HOSTNAME environment variable only works with MQ clients 9.2.1 or newer.
Depending on the container runtime environment you are using, if you have problems creating enough application (and MQ client threads) in the cphtestp container, you may need to increase the PID limit. This can be configured on the docker/podman run --pids-limit=8192 cphtestp or more permanently in the /usr/share/containers/containers.conf file.
The MQ clients within the container can be configured to attempt to autoreconnect, by setting the MQ_AUTORECONNECT environment variable. This can be set to any of the MQ CNO reconnection options. Setting it to MQCNO_RECONNECT - the clients will attempt to reconnect using the same QMName, host and port configuration as the initial connection. By default, the functionality will remain MQCNO_RECONNECT_DISABLED. It should be noted that the underlying MQ client (cph) is not coded to support reliable reconnection in all circumstances.
The version of cph contained in this image was taken on 3rd July 2024 and built on 64bit xLinux. The most up to date cph code can be found here: https://github.com/ibm-messaging/mq-cph
Set MQ_COMPRESS to TRUE to enable all client compression algorithms at the client. The specific compression algorithm to use can then be controlled by the QM CHANNEL configuration. We would recommend LZ4FAST in scenarios where you may be network restricted.