Performance monitoring and testing of OpenStack
Go to file
Sai Sindhur Malleni 5a48323531 Adding Glance Rally Scenarios
Adding some common scenarios for stressing glance.
+ Create and list images
+ Create and delete images
+ Create image and boot instances off it
+ List images by placing image in context

Also adding a separate configuration file for glance only.

Change-Id: I5327e9c65cb1b045f686e442e88b3e71f73ac5cf
2016-08-01 23:50:55 +00:00
ansible Fix role ordering in collectd-openstack.yml 2016-08-01 18:56:18 +00:00
conf Adding Glance Rally Scenarios 2016-08-01 23:50:55 +00:00
docs/source Adding Browbeat Infra files 2016-06-14 09:16:53 -04:00
lib Merge "Handle case where benchmark list is empty" 2016-07-22 17:46:18 +00:00
log Install updates. 2016-02-15 16:25:10 +01:00
metadata Updates to metadata 2016-06-16 15:53:11 -04:00
rally Adding Glance Rally Scenarios 2016-08-01 23:50:55 +00:00
results Install updates. 2016-02-15 16:25:10 +01:00
utils Add keystone get-token.py utility 2016-07-14 08:14:31 -04:00
.gitignore Fix issues with Keystone token adjustment playbook 2016-06-23 08:51:28 -04:00
.gitreview Adding Browbeat Infra files 2016-06-14 09:16:53 -04:00
LICENSE Add LICENSE file for Apache 2.0 license 2016-05-06 11:25:17 -04:00
README.rst Remove TripleO as a req 2016-07-10 20:56:02 -04:00
browbeat-complete.yaml Adding Glance Rally Scenarios 2016-08-01 23:50:55 +00:00
browbeat-config.yaml Adding Glance Rally Scenarios 2016-08-01 23:50:55 +00:00
browbeat.py Index each document with a run UUID. 2016-06-20 09:38:49 -04:00
requirements.txt Adding Browbeat Infra files 2016-06-14 09:16:53 -04:00
setup.cfg Adding Browbeat Infra files 2016-06-14 09:16:53 -04:00
setup.py Adding Browbeat Infra files 2016-06-14 09:16:53 -04:00
test-requirements.txt Adding Browbeat Infra files 2016-06-14 09:16:53 -04:00
tox.ini Adding Browbeat Infra files 2016-06-14 09:16:53 -04:00

README.rst

Table of Contents

Browbeat

This started as a project to help determine the number of database connections a given OpenStack deployment uses via stress tests. It has since grown into a set of Ansible playbooks to help check deployments for known issues, install tools, run performance stress workloads and change parameters of the overcloud.

Before running browbeat

  • Execute the ansible/generate_tripleo_hostfile.sh script (builds the ssh config)
  • Configure browbeat-config.yaml to match your tests

How to run Browbeat Stress Tests?

On the Undercloud host, as the Stack user jump into the browbeat venv and you simply run:

(browbeat-venv)[stack@ospd browbeat]$ ./browbeat.py --help

However, the playbook required to install browbeat and its dependencies(Rally, Shaker, Perfkit) needs to be run before this. DEtailed install and run instructions are presented in a section below.

What is necessary?

  • Ansible

    Why? We started with using bash to make changes to the Overcloud, creating complex sed/awks that we get for free with Ansible (for the most part). Other monitoring and stress test tools are installed by the respective playbooks when run.

Detailed Install, Check and Run

Installing Browbeat and running the Overcloud checks can be performed either from your local machine or from the undercloud. The local machine install/check assumes you have ansible installed already.

Install Browbeat from your local machine

From your local machine

$ ssh-copy-id stack@<undercloud-ip>
$ git clone https://github.com/openstack/browbeat.git
$ cd browbeat/ansible
$ ./generate_tripleo_hostfile.sh <undercloud-ip> ~/.ssh/config
$ vi install/group_vars/all.yml # Make sure to edit the dns_server to the correct ip address
$ ansible-playbook -i hosts install/browbeat.yml
$ vi install/group_vars/all.yml # Edit browbeat network settings
$ ansible-playbook -i hosts install/browbeat_network.yml
$ ansible-playbook -i hosts install/shaker_build.yml

(Optional) Install collectd

$ ansible-playbook -i hosts install/collectd-openstack.yml

(Optional) Install collectd->graphite dashboards

$ ansible-playbook -i hosts install/dashboards-openstack.yml

(Optional) Install connmon

$ ansible-playbook -i hosts install/connmon.yml

Run Overcloud checks

$ ansible-playbook -i hosts check/site.yml

Your Overcloud check output is located in check/bug_report.log

NOTE: It is strongly advised to not run the ansible playbooks in a venv.

Run performance stress tests through browbeat on the undercloud:

$ ssh undercloud-root
[root@ospd ~]# su - stack
[stack@ospd ~]$ screen -S browbeat
[stack@ospd ~]$ . browbeat-venv/bin/activate
(browbeat-venv)[stack@ospd ~]$ cd browbeat/
(browbeat-venv)[stack@ospd browbeat]$ vi browbeat-config.yaml # Edit browbeat-config.yaml to control how many stress tests are run.
(browbeat-venv)[stack@ospd browbeat]$ ./browbeat.py <workload> #perfkit, rally, shaker or "all"

Install Browbeat directly on undercloud

From your undercloud

$ ssh undercloud-root
[root@ospd ~]# su - stack
[stack@ospd ~]$ git clone https://github.com/openstack/browbeat.git
[stack@ospd ~]$ cd browbeat/ansible
[stack@ospd ansible]$ ./generate_tripleo_hostfile.sh localhost ~/.ssh/config
[stack@ospd ansible]$ sudo easy_install pip
[stack@ospd ansible]$ sudo pip install ansible
[stack@ospd ansible]$ vi install/group_vars/all.yml # Make sure to edit the dns_server to the correct ip address
[stack@ospd ansible]$ ansible-playbook -i hosts install/browbeat.yml
[stack@ospd ansible]$ vi install/group_vars/all.yml # Edit browbeat network settings
[stack@ospd ansible]$ ansible-playbook -i hosts install/browbeat_network.yml
[stack@ospd ansible]$ ansible-playbook -i hosts install/shaker_build.yml

(Optional) Install collectd

[stack@ospd ansible]$ ansible-playbook -i hosts install/collectd-openstack.yml

(Optional) Install collectd->graphite dashboards

[stack@ospd ansible]$ ansible-playbook -i hosts install/dashboards-openstack.yml

(Optional) Install connmon

[stack@ospd ansible]$ ansible-playbook -i hosts install/connmon.yml

Run Overcloud checks

[stack@ospd ansible]$ ansible-playbook -i hosts check/site.yml

Your Overcloud check output is located in check/bug_report.log

Run performance stress tests through browbeat

[stack@ospd ansible]$ . ../../browbeat-venv/bin/activate
(browbeat-venv)[stack@ospd ansible]$ cd ..
(browbeat-venv)[stack@ospd browbeat]$ vi browbeat-config.yaml # Edit browbeat.cfg to control how many stress tests are run.
(browbeat-venv)[stack@ospd browbeat]$ ./browbeat.py <workload> #perfkit, rally, shaker or "all"

Running PerfKitBenchmarker

Work is on-going to utilize PerfKitBenchmarker as a workload provider to browbeat. Many benchmarks work out of the box with browbeat. You must ensure that your network is setup correctly to run those benchmarks and you will need to configure the settings in ansible/install/group_vars/all.yml for browbeat public/private networks. Currently tested benchmarks include: aerospike, bonnie++, cluster_boot, copy_throughput(cp,dd,scp), fio, iperf, mesh_network, mongodb_ycsb, netperf, object_storage_service, ping, scimark2, and sysbench_oltp.

To run browbeat's PerfKit Benchmarks, you can start by viewing the tested benchmark's configuration in conf/browbeat-perfkit-complete.yaml. You must add them to your specific browbeat config yaml file or enable/disable the benchmarks you wish to run in the default config file (browbeat-config.yaml). There are many flags exposed in the configuration files to tune how those benchmarks run. Additional flags are exposed in the source code of PerfKitBenchmarker available on the Google Cloud Github.

Example running only PerfKitBenchmarker benchmarks with browbeat from browbeat-config.yaml:

(browbeat-venv)[stack@ospd browbeat]$ ./browbeat.py  perfkit -s browbeat-config.yaml

Contributing

Contributions are most welcome! You must first create a Launchpad account and follow the instructions here to get started as a new OpenStack contributor.

Once you've signed the contributor license agreement and read through the above documentation, add your public SSH key under the 'SSH Public Keys' section of review.openstack.org.

You can view your public key using:

$ cat ~/.ssh/id_*.pub

Set your username and email for review.openstack.org:

$ git config --global user.email "example@example.com"
$ git config --global user.name "example"
$ git config --global --add gitreview.username "example"

Next, Clone the github repository:

$ git clone https://github.com/openstack/browbeat.git

You need to have git-review in order to be able to submit patches using the gerrit code review system. You can install it using:

$ sudo yum install git-review

To set up your cloned repository to work with OpenStack Gerrit

$ git review -s

It's useful to create a branch to do your work, name it something related to the change you'd like to introduce.

$ cd browbeat
$ git branch my_special_enhancement
$ git checkout !$

Make your changes and then commit them using the instructions below.

$ git add /path/to/files/changed
$ git commit

Use a descriptive commit title followed by an empty space. You should type a small justification of what you are changing and why.

Now you're ready to submit your changes for review:

$ git review

If you want to make another patchset from the same commit you can use the ammend feature after further modification and saving.

$ git add /path/to/files/changed
$ git commit --amend
$ git review

If you want to submit a new patchset from a different location (perhaps on a different machine or computer for example) you can clone the browbeat repo again (if it doesn't already exist) and then use git review against your unique Change-ID:

$ git review -d Change-Id

Change-Id is the change id number as seen in Gerrit and will be generated after your first successful submission.

The above command downloads your patch onto a seperate branch. You might need to rebase your local branch with remote master before running it to avoid merge conflicts when you resubmit the edited patch. To avoid this go back to a "safe" commit using:

$ git reset --hard commit-number

Then,

$ git fetch origin
$ git rebase origin/master

Make the changes on the branch that was setup by using the git review -d (the name of the branch is along the lines of review/username/branch_name/patchsetnumber).

Add the files to git and commit your changes using,

$ git commit --amend

You can edit your commit message as well in the prompt shown upon executing above command.

Finally, push the patch for review using,

$ git review

Resources