Merge "Fix formatting + add some details to k8s api testing results"

This commit is contained in:
Jenkins 2016-10-07 06:28:08 +00:00 committed by Gerrit Code Review
commit f68133ee13

View File

@ -69,8 +69,8 @@ Network scheme of the environment:
:alt: Network Scheme of the environment
:scale: 65
Here is the piece of switch configuration for each switch port which is a part of
bond0 interface of a server:
Here is the piece of switch configuration for each switch port which is a part
of bond0 interface of a server:
.. code:: bash
@ -155,9 +155,10 @@ to nodes count):
Operating system configuration
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
You can find outputs of some commands and /etc folder in the following archive:
:download:`server_description_of_node1 <configs/server_description_of_node1.tar.gz>`
:download:`server_description_of_node3 <configs/server_description_of_node3.tar.gz>`
:download:`server_description_of_node4 <configs/server_description_of_node4.tar.gz>`
- :download:`server_description_of_node1 <configs/server_description_of_node1.tar.gz>`
- :download:`server_description_of_node3 <configs/server_description_of_node3.tar.gz>`
- :download:`server_description_of_node4 <configs/server_description_of_node4.tar.gz>`
Software configuration of Test tool:
------------------------------------
@ -251,18 +252,18 @@ We performed the steps from 1 to 3 for Kubernetes cluster on top of 10, 50 and
Results
=======
10-nodes cluster
----------------
10-nodes cluster (all values are presented in milliseconds)
-----------------------------------------------------------
.. include:: results/10_nodes_results.rst
50-nodes cluster
----------------
50-nodes cluster (all values are presented in milliseconds)
-----------------------------------------------------------
.. include:: results/50_nodes_results.rst
355-nodes cluster
-----------------
355-nodes cluster (all values are presented in milliseconds)
------------------------------------------------------------
.. include:: results/355_nodes_results.rst
@ -299,13 +300,15 @@ For this testing purposes `MMM(MySQL/Master/Minions) testing suite`_ was used
(more information in `Pod startup time measurement toolkit`_ section).
This toolkit was run against 150 nodes Kubernetes environment installed via
`Kargo`_ deployment tool. The most basic configuration (1 replication
controller, N pods, each pod containing 1 container) was run against the
environment. Additional configurations will be tested and results published
in terms of further researches.
`Kargo`_ deployment tool (these nodes were taken from the same nodes pool all
previous Kubernetes API performance tests were run against). The most basic
configuration (1 replication controller, N pods, each pod containing 1
container) was run against the environment. Additional configurations will be
tested and results published in terms of further researches.
The first run includes information about 500 pods being run on fresh Kubernetes
environment (no tests have been run on it before):
environment (no tests have been run on it before, warm-up run with about 3 pods
per node density):
.. image:: MMM-tool/500_containers_first_run.png
:alt: Containers startup time (500 containers, first run)
@ -317,14 +320,16 @@ pre-loaded on Kubernetes worker nodes, that means that during first run Docker
registry/repo/etc was really stressed).
The same scenario run against the warmed-up environment will have linear
pattern (with ~50 milliseconds per container startup):
pattern (with ~50 milliseconds per container startup, about 3 pods per cluster
node density):
.. image:: MMM-tool/500_containers_second_run.png
:alt: Containers startup time (500 containers, second run)
:width: 650px
This pattern will remain the same with bigger number of containers (15000
containers, the same ~50 milliseconds per container startup):
containers, the same ~50 milliseconds per container startup, 100 pods per
cluster node density):
.. image:: MMM-tool/15000_containers.png
:alt: Containers startup time (15000 containers)