openstack/performance-docs
Code Issues Proposed changes

Add reliability test results

Browse Source 
This commit add part of reliability testing results.
Scope of this commit is testing Nova API under
several factors.

Change-Id: Id3cb644ccf4bd315846399e6ac40a446297787f3
This commit is contained in:
Alexandr Nevenchannyy 2016-06-23 19:43:20 +03:00
parent 92816644f7
commit f84ec2ce07
21 changed files with 6428 additions and 37 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

38
doc/source/test_plans/reliability/plan.rst
View File

@ -14,43 +14,7 @@ OpenStack reliability testing
:Conventions: :Conventions:
- **OpenStack cluster:** consists of server nodes with deployed and fully .. include:: plan_conventions.rst
operational OpenStack environment in high-availability configuration.
- **Fault-injection operation:** represents common types of failures which can
occur in production environment: service-hang, service-crash,
network-partition, network-flapping, and node-crash.
- **Service-hang:** faults are injected into specified OpenStack service by
sending -SIGSTOP and -SIGCONT POSIX signals.
- **Service-crash:** faults are injected by sending -SIGKILL signal into
specified OpenStack service.
- **Node-crash:** faults are injected to an OpenStack cluster by rebooting
or shutting down a server node.
- **Network-partition:** faults are injected by inserting iptables rules to
OpenStack cluster nodes to a corresponding service that should be
network-partitioned.
- **Network-flapping:** faults are injected into OpenStack cluster nodes by
inserting/deleting iptables rules on the fly which will affect
corresponding service that should be tested.
- **Factor:** consists of a set of atomic fault-injection operations. For
example: reboot-random-controller, reboot-random-rabbitmq.
- **Test plan:** contains two elements: test scenario
execution graph and fault-injection factors.
- **SLA**: Service-level agreement
- **Testing-cycles**: number of test cycles of each factor
- **Inf**: assumes infinite time to auto-healing of cluster
after fault-factor injection.
Test Plan Test Plan
========= =========

36
doc/source/test_plans/reliability/plan_conventions.rst Normal file
View File

@ -0,0 +1,36 @@
- **OpenStack cluster:** consists of server nodes with deployed and fully
operational OpenStack environment in high-availability configuration.
- **Fault-injection operation:** represents common types of failures which can
occur in production environment: service-hang, service-crash,
network-partition, network-flapping, and node-crash.
- **Service-hang:** faults are injected into specified OpenStack service by
sending -SIGSTOP and -SIGCONT POSIX signals.
- **Service-crash:** faults are injected by sending -SIGKILL signal into
specified OpenStack service.
- **Node-crash:** faults are injected to an OpenStack cluster by rebooting
or shutting down a server node.
- **Network-partition:** faults are injected by inserting iptables rules to
OpenStack cluster nodes to a corresponding service that should be
network-partitioned.
- **Network-flapping:** faults are injected into OpenStack cluster nodes by
inserting/deleting iptables rules on the fly which will affect
corresponding service that should be tested.
- **Factor:** consists of a set of atomic fault-injection operations. For
example: reboot-random-controller, reboot-random-rabbitmq.
- **Test plan:** contains two elements: test scenario
execution graph and fault-injection factors.
- **SLA**: Service-level agreement
- **Testing-cycles**: number of test cycles of each factor
- **Inf**: assumes infinite time to auto-healing of cluster
after fault-factor injection.

1
doc/source/test_results/index.rst
View File

@ -20,3 +20,4 @@ Test Results
hardware_features/index hardware_features/index
provisioning/index provisioning/index
1000_nodes/index 1000_nodes/index
reliability/index

BIN
doc/source/test_results/reliability/images/Network_Scheme.png Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 20 KiB

650
doc/source/test_results/reliability/index.rst Normal file
View File

@ -0,0 +1,650 @@
.. _reliability_testing_results:
=============================
OpenStack reliability testing
=============================
:status: draft
:version: 0
:Abstract:
This document describes an abstract methodology for OpenStack cluster
high-availability testing and analysis. OpenStack data plane testing
at this moment is out of scope but will be described in future.
:Conventions:
.. include:: ../../test_plans/reliability/plan_conventions.rst
Test results
============
Test environment
----------------
Software configuration on servers with OpenStack
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
.. table:: **Basic cluster configuration**
+-------------------------+---------------------------------------------+
|Name | Build-9.0.0-451 |
+-------------------------+---------------------------------------------+
|OpenStack release | Mitaka on Ubuntu 14.04 |
+-------------------------+---------------------------------------------+
|Total nodes | 6 nodes |
+-------------------------+---------------------------------------------+
|Controller | 3 nodes |
+-------------------------+---------------------------------------------+
|Compute, Ceph OSD | 3 nodes with KVM hypervisor |
+-------------------------+---------------------------------------------+
|Network | Neutron with tunneling segmentation |
+-------------------------+---------------------------------------------+
|Storage back ends | | Ceph RBD for volumes (Cinder) |
| | | Ceph RadosGW for objects (Swift API) |
| | | Ceph RBD for ephemeral volumes (Nova) |
| | | Ceph RBD for images (Glance) |
+-------------------------+---------------------------------------------+
Software configuration on servers with Rally role
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Before you start configuring a server with Rally role, verify that Rally
is installed. For more information, see `Rally installation documentation`_.
.. table:: **Software version of Rally server**
+------------+-------------------+
|Software |Version |
+============+===================+
|Rally |0.4.0 |
+------------+-------------------+
|Ubuntu |14.04.3 LTS |
+------------+-------------------+
Environment description
^^^^^^^^^^^^^^^^^^^^^^^
Hardware
~~~~~~~~
.. table:: **Description of server hardware**
+--------+-----------------+------------------------+------------------------+
|SERVER |name | | 728997-comp-disk-228 | 729017-comp-disk-255 |
| | | | 728998-comp-disk-227 | |
| | | | 728999-comp-disk-226 | |
| | | | 729000-comp-disk-225 | |
| | | | 729001-comp-disk-224 | |
| | | | 729002-comp-disk-223 | |
| +-----------------+------------------------+------------------------+
| |role | | controller | Rally |
| | | | controller | |
| | | | controller | |
| | | | compute, ceph-osd | |
| | | | compute, ceph-osd | |
| | | | compute, ceph-osd | |
| +-----------------+------------------------+------------------------+
| |vendor, model |HP, DL380 Gen9 |HP, DL380 Gen9 |
| +-----------------+------------------------+------------------------+
| |operating_system | | 3.13.0-87-generic | | 3.13.0-87-generic |
| | | | Ubuntu-trusty | | Ubuntu-trusty |
| | | | x86_64 | | x86_64 |
+--------+-----------------+------------------------+------------------------+
|CPU |vendor, model |Intel, E5-2680 v3 |Intel, E5-2680 v3 |
| +-----------------+------------------------+------------------------+
| |processor_count |2 |2 |
| +-----------------+------------------------+------------------------+
| |core_count |12 |12 |
| +-----------------+------------------------+------------------------+
| |frequency_MHz |2500 |2500 |
+--------+-----------------+------------------------+------------------------+
|RAM |vendor, model |HP, 752369-081 |HP, 752369-081 |
| +-----------------+------------------------+------------------------+
| |amount_MB |262144 |262144 |
+--------+-----------------+------------------------+------------------------+
|NETWORK |interface_name |p1p1 |p1p1 |
| +-----------------+------------------------+------------------------+
| |vendor, model |Intel, X710 Dual Port |Intel, X710 Dual Port |
| +-----------------+------------------------+------------------------+
| |bandwidth |10 Gbit |10 Gbit |
+--------+-----------------+------------------------+------------------------+
|STORAGE |dev_name |/dev/sda |/dev/sda |
| +-----------------+------------------------+------------------------+
| |vendor, model | | raid10 - HP P840 | | raid10 - HP P840 |
| | | | 12 disks EH0600JEDHE | | 12 disks EH0600JEDHE |
| +-----------------+------------------------+------------------------+
| |SSD/HDD |HDD |HDD |
| +-----------------+------------------------+------------------------+
| |size | 3,6 TB | 3,6 TB |
+--------+-----------------+------------------------+------------------------+
Software
~~~~~~~~
.. table:: **Services on servers by role**
+------------+----------------------------+
|Role |Service name |
+============+============================+
|controller || horizon |
| || keystone |
| || nova-api |
| || nava-scheduler |
| || nova-cert |
| || nova-conductor |
| || nova-consoleauth |
| || nova-consoleproxy |
| || cinder-api |
| || cinder-backup |
| || cinder-scheduler |
| || cinder-volume |
| || glance-api |
| || glance-glare |
| || glance-registry |
| || neutron-dhcp-agent |
| || neutron-l3-agent |
| || neutron-metadata-agent |
| || neutron-openvswitch-agent |
| || neutron-server |
| || heat-api |
| || heat-api-cfn |
| || heat-api-cloudwatch |
| || ceph-mon |
| || rados-gw |
| || heat-engine |
| || memcached |
| || rabbitmq_server |
| || mysqld |
| || galera |
| || corosync |
| || pacemaker |
| || haproxy |
+------------+----------------------------+
|compute-osd || nova-compute |
| || neutron-l3-agent |
| || neutron-metadata-agent |
| || neutron-openvswitch-agent |
| || ceph-osd |
+------------+----------------------------+
High availability cluster architecture
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Controller nodes:
.. image:: https://docs.mirantis.com/openstack/fuel/fuel-8.0/_images/logical-diagram-controller.svg
:height: 700px
:width: 600px
:alt: Mirantis reference HA architecture
Compute nodes:
.. image:: https://docs.mirantis.com/openstack/fuel/fuel-8.0/_images/logical-diagram-compute.svg
:height: 250px
:width: 350px
:alt: Mirantis reference HA architecture
Networking
~~~~~~~~~~
All servers have the similar network configuration:
.. image:: images/Network_Scheme.png
:alt: Network Scheme of the environment
The following example shows a part of a switch configuration for each switch
port that is connected to ens1f0 interface of a server:
.. code:: bash
switchport mode trunk
switchport trunk native vlan 600
switchport trunk allowed vlan 600-602,630-649
spanning-tree port type edge trunk
spanning-tree bpduguard enable
no snmp trap link-status
Factors description
-------------------
- **reboot-random-controller:** consists of a node-crash fault injection on a
random OpenStack controller node.
- **sigkill-random-rabbitmq:** consists of a service-crash fault injection on
a random slave RabbitMQ messaging node.
- **sigkill-random-mysql:** consists of a service-crash fault injection on a
random MySQL node.
- **freeze-random-nova-api:** consists of a service-hang fault injection to
all nova-api process on a random controller node for a 150 seconds period.
- **freeze-random-memcached:** consists of a service-hang fault injection to
the memcached service on a random controller node for a 150 seconds period.
- **freeze-random-keystone:** consists of a service-hang fault injection to
the keystone (public and admin endpoints) service on a random controller
node for a 150 seconds period.
Testing process
===============
Use the following VM parameters for testing purposes:
.. table:: **Test parameters**
+--------------------------------+--------+
|Name |Value |
+================================+========+
|Flavor to create VM from |m1.tiny |
+--------------------------------+--------+
|Image name to create VM from |cirros |
+--------------------------------+--------+
#. Create a work directory on a server with Rally role.
In this documentation, we name this directory ``WORK_DIR``. The path
example: ``/data/rally``.
#. Create a directory ``plugins`` in ``WORK_DIR`` and copy the
:download:`scrappy.py <rally_plugins/scrappy.py>` plugin into that directory.
#. Download the bash framework :download:`scrappy.sh <rally_plugins/scrappy.sh>`
and :download:`scrappy.conf <rally_plugins/scrappy.conf>` to
``WORK_DIR/plugins``.
#. Modify the ``scrappy.conf`` file with appropriate values. For example:
.. literalinclude:: rally_plugins/scrappy.conf
:language: bash
#. Create a ``scenarios`` directory in ``WORK_DIR`` and copy all Rally
scenarios with factors that you are planning to test to that directory.
For example:
:download:`random_controller_reboot_factor.json
<rally_scenarios/NovaServers/boot_and_delete_server/random_controller_reboot_factor.json/>`.
#. Create a ``deployment.json`` file in the ``WORK_DIR`` and fill it with
your OpenStack environment info. It should looks like this:
.. code:: json
{
"admin": {
"password": "password",
"tenant_name": "tenant",
"username": "user"
},
"auth_url": "http://1.2.3.4:5000/v2.0",
"region_name": "RegionOne",
"type": "ExistingCloud",
"endpoint_type": "internal",
"admin_port": 35357,
"https_insecure": true
}
#. Prepare for tests:
.. code:: bash
${WORK_DIR:?}
DEPLOYMENT_NAME="$(uuidgen)"
DEPLOYMENT_CONFIG="${WORK_DIR}/deployment.json"
rally deployment create --filename $(DEPLOYMENT_CONFIG) --name $(DEPLOYMENT_NAME)
#. Create a ``/root/scrappy`` directory on every node in your OpenStack
environment and copy :download:`scrappy_host.sh <rally_plugins/scrappy_host.sh>`
to that directory.
#. Perform tests:
.. code:: bash
PLUGIN_PATH="${WORK_DIR}/plugins"
SCENARIOS="random_controller_reboot_factor.json"
for scenario in SCENARIOS; do
rally --plugin-paths ${PLUGINS_PATH} task start --tag ${scenario} ${WORK_DR}/scenarios/${scenario}
done
task_list="$(rally task list --uuids-only)"
rally task report --tasks ${task_list} --out=${WORK_DIR}/rally_report.html
Once these steps are done, you get an HTML file with Rally test results.
Test case 1: NovaServers.boot_and_delete_server
-----------------------------------------------
**Description**
This Rally scenario boots and deletes virtual instances with injected fault
factors using OpenStack Nova API.
**Service-level agreement**
=================== ========
Parameter Value
=================== ========
MTTR (sec) <=240
Failure rate (%) <=95
Auto-healing Yes
=================== ========
**Parameters**
=================== ========
Parameter Value
=================== ========
Runner constant
Concurrency 5
Times 100
Injection-iteration 20
Testing-cycles 5
=================== ========
**List of reliability metrics**
======== ============== ================= =================================================
Priority Value Measurement Units Description
======== ============== ================= =================================================
1 SLA Boolean Service-level agreement result
2 Auto-healing Boolean Is cluster auto-healed after fault-injection
3 Failure rate Percents Test iteration failure ratio
4 MTTR (auto) Seconds Automatic mean time to repair
5 MTTR (manual) Seconds Manual mean time to repair, if Auto MTTR is Inf.
======== ============== ================= =================================================
Test case 1 results
-------------------
reboot-random-controller
~~~~~~~~~~~~~~~~~~~~~~~~
**Rally scenario used during factor testing:**
.. literalinclude:: rally_scenarios/NovaServers/boot_and_delete_server/random_controller_reboot_factor.json
:language: bash
**Factor testing results:**
.. table:: **Full description of cyclic execution results**
+--------+-----------+-----------------+--------------+-------------------------+
| Cycles | MTTR(sec) | Failure rate(%) | Auto-healing | Performance degradation |
+--------+-----------+-----------------+--------------+-------------------------+
| 1 | 4.31 | 2 | Yes | Yes, up to 148.52 sec. |
+--------+-----------+-----------------+--------------+-------------------------+
| 2 | 19.88 | 14 | Yes | Yes, up to 150.946 sec. |
+--------+-----------+-----------------+--------------+-------------------------+
| 3 | 7.31 | 8 | Yes | Yes, up to 124.593 sec. |
+--------+-----------+-----------------+--------------+-------------------------+
| 4 | 95.07 | 9 | Yes | Yes, up to 240.893 |
+--------+-----------+-----------------+--------------+-------------------------+
| 5 | Inf. | 80.00 | No | Inf. |
+--------+-----------+-----------------+--------------+-------------------------+
**Rally report:** :download:`reboot_random_controller.html <../../../../raw_results/reliability/rally_results/NovaServers/boot_and_delete_server/reboot_random_controller.html>`
.. table:: **Testing results summary**
+---------------+-----------------+------------------+
| Value | MTTR(sec) | Failure rate |
+---------------+-----------------+------------------+
| Min | 4.31 | 2 |
+---------------+-----------------+------------------+
| Max | 95.07 | 80 |
+---------------+-----------------+------------------+
| SLA | Yes | No |
+---------------+-----------------+------------------+
**Detailed results description**
This factor affects OpenStack cluster operation on every run.
Auto-healing works, but may take a long time. In our testing results, the
cluster was recovered on the fifth testing cycle only, after Rally had
completed testing with the error status. Therefore, the performance degradation
is very significant during cluster recovering.
sigkill-random-rabbitmq
~~~~~~~~~~~~~~~~~~~~~~~
**Rally scenario used during factor testing:**
.. literalinclude:: rally_scenarios/NovaServers/boot_and_delete_server/random_controller_kill_rabbitmq.json
:language: bash
**Factor testing results:**
.. table:: **Full description of cyclic execution results**
+--------------------+----------------+---------------------+------------------+-----------------------------+
| Cycles | MTTR(sec) | Failure rate(%) | Auto-healing | Performance degradation |
+--------------------+----------------+---------------------+------------------+-----------------------------+
| 1 | 0 | 0 | Yes | Yes, up to 12.266 sec. |
+--------------------+----------------+---------------------+------------------+-----------------------------+
| 2 | 0 | 0 | Yes | Yes, up to 15.775 sec. |
+--------------------+----------------+---------------------+------------------+-----------------------------+
| 3 | 98.52 | 1 | Yes | Yes, up to 145.115 sec. |
+--------------------+----------------+---------------------+------------------+-----------------------------+
| 4 | 0 | 0 | Yes | No |
+--------------------+----------------+---------------------+------------------+-----------------------------+
| 5 | 0 | 0 | Yes | Yes, up to 65.926 sec. |
+--------------------+----------------+---------------------+------------------+-----------------------------+
**Rally report:** :download:`random_controller_kill_rabbitmq.html <../../../../raw_results/reliability/rally_results/NovaServers/boot_and_delete_server/random_controller_kill_rabbitmq.html>`
.. table:: **Testing results summary**
+--------+-----------+--------------+
| Value | MTTR(sec) | Failure rate |
+--------+-----------+--------------+
| Min | 0 | 0 |
+--------+-----------+--------------+
| Max | 98.52 | 1 |
+--------+-----------+--------------+
| SLA | Yes | Yes |
+--------+-----------+--------------+
**Detailed results description**
This factor may affect OpenStack cluster operation.
Auto-healing works fine.
Performance degradation is significant during cluster recovering.
sigkill-random-mysql
~~~~~~~~~~~~~~~~~~~~
**Rally scenario used during factor testing:**
.. literalinclude:: rally_scenarios/NovaServers/boot_and_delete_server/random_controller_kill_mysqld.json
:language: bash
**Factor testing results:**
.. table:: **Full description of cyclic execution results**
+--------------------+----------------+---------------------+------------------+-----------------------------+
| Cycles | MTTR(sec) | Failure rate(%) | Auto-healing | Performance degradation |
+--------------------+----------------+---------------------+------------------+-----------------------------+
| 1 | 2.31 | 0 | Yes | Yes, up to 12.928 sec. |
+--------------------+----------------+---------------------+------------------+-----------------------------+
| 2 | 0 | 0 | Yes | Yes, up to 11.156 sec. |
+--------------------+----------------+---------------------+------------------+-----------------------------+
| 3 | 0 | 1 | Yes | Yes, up to 13.592 sec. |
+--------------------+----------------+---------------------+------------------+-----------------------------+
| 4 | 0 | 0 | Yes | Yes, up to 11.864 sec. |
+--------------------+----------------+---------------------+------------------+-----------------------------+
| 5 | 0 | 0 | Yes | Yes, up to 12.715 sec. |
+--------------------+----------------+---------------------+------------------+-----------------------------+
**Rally report:** :download:`random_controller_kill_mysqld.html <../../../../raw_results/reliability/rally_results/NovaServers/boot_and_delete_server/random_controller_kill_mysqld.html>`
.. table:: **Testing results summary**
+--------+-----------+--------------+
| Value | MTTR(sec) | Failure rate |
+--------+-----------+--------------+
| Min | 0 | 0 |
+--------+-----------+--------------+
| Max | 2.31 | 1 |
+--------+-----------+--------------+
| SLA | Yes | Yes |
+--------+-----------+--------------+
**Detailed results description**
This factor may affect OpenStack cluster operation.
Auto-healing works fine.
Performance degradation is not significant.
freeze-random-nova-api
~~~~~~~~~~~~~~~~~~~~~~
**Rally scenario used during factor testing:**
.. literalinclude:: rally_scenarios/NovaServers/boot_and_delete_server/random_controller_freeze_nova-api_150_sec.json
:language: bash
**Factor testing results:**
.. table:: **Full description of cyclic execution results**
+--------------------+----------------+---------------------+------------------+-----------------------------+
| Cycles | MTTR(sec) | Failure rate(%) | Auto-healing | Performance degradation |
+--------------------+----------------+---------------------+------------------+-----------------------------+
| 1 | 0 | 0 | Yes | Yes, up to 156.935 sec. |
+--------------------+----------------+---------------------+------------------+-----------------------------+
| 2 | 0 | 0 | Yes | Yes, up to 155.085 sec. |
+--------------------+----------------+---------------------+------------------+-----------------------------+
| 3 | 0 | 0 | Yes | Yes, up to 156.93 sec. |
+--------------------+----------------+---------------------+------------------+-----------------------------+
| 4 | 0 | 0 | Yes | Yes, up to 156.782 sec. |
+--------------------+----------------+---------------------+------------------+-----------------------------+
| 5 | 150.55 | 1 | Yes | Yes, up to 154.741 sec. |
+--------------------+----------------+---------------------+------------------+-----------------------------+
**Rally report:** :download:`random_controller_freeze_nova_api_150_sec.html <../../../../raw_results/reliability/rally_results/NovaServers/boot_and_delete_server/random_controller_freeze_nova_api_150_sec.html>`
.. table:: **Testing results summary**
+--------+-----------+--------------+
| Value | MTTR(sec) | Failure rate |
+--------+-----------+--------------+
| Min | 0 | 0 |
+--------+-----------+--------------+
| Max | 150.55 | 1 |
+--------+-----------+--------------+
| SLA | Yes | Yes |
+--------+-----------+--------------+
**Detailed results description**
This factor affects OpenStack cluster operation.
Auto-healing does not work. Cluster operation was recovered
only after sending SIGCONT POSIX signal to all freezed nova-api
processes. Performance degradation is determined by the factor duration time.
This behaviour is not normal for an HA OpenStack configuration
and should be investigated.
freeze-random-memcached
~~~~~~~~~~~~~~~~~~~~~~~
**Rally scenario used during factor testing:**
.. literalinclude:: rally_scenarios/NovaServers/boot_and_delete_server/random_controller_freeze_memcached_150_sec.json
:language: bash
**Factor testing results:**
.. table:: **Full description of cyclic execution results**
+--------------------+----------------+---------------------+------------------+-----------------------------+
| Cycles | MTTR(sec) | Failure rate(%) | Auto-healing | Performance degradation |
+--------------------+----------------+---------------------+------------------+-----------------------------+
| 1 | 0 | 0 | Yes | Yes, up to 26.679 sec. |
+--------------------+----------------+---------------------+------------------+-----------------------------+
| 2 | 0 | 0 | Yes | Yes, up to 23.726 sec. |
+--------------------+----------------+---------------------+------------------+-----------------------------+
| 3 | 0 | 0 | Yes | Yes, up to 21.893 sec. |
+--------------------+----------------+---------------------+------------------+-----------------------------+
| 4 | 0 | 0 | Yes | Yes, up to 22.796 sec. |
+--------------------+----------------+---------------------+------------------+-----------------------------+
| 5 | 0 | 0 | Yes | Yes, up to 27.737 sec. |
+--------------------+----------------+---------------------+------------------+-----------------------------+
**Rally report:** :download:`random_controller_freeze_memcached_150_sec.html <../../../../raw_results/reliability/rally_results/NovaServers/boot_and_delete_server/random_controller_freeze_memcached_150_sec.html>`
.. table:: **Testing results summary**
+--------+-----------+--------------+
| Value | MTTR(sec) | Failure rate |
+--------+-----------+--------------+
| Min | 0 | 0 |
+--------+-----------+--------------+
| Max | 0 | 0 |
+--------+-----------+--------------+
| SLA | Yes | Yes |
+--------+-----------+--------------+
**Detailed results description**
This factor does not affect an OpenStack cluster operations.
During the factor testing, a small performance degradation is observed.
freeze-random-keystone
~~~~~~~~~~~~~~~~~~~~~~
**Rally scenario used during factor testing:**
.. literalinclude:: rally_scenarios/NovaServers/boot_and_delete_server/random_controller_freeze_keystone_150_sec.json
:language: bash
**Factor testing results:**
.. table:: **Full description of cyclic execution results**
+--------+-------------+-----------------+--------------+-------------------------+
| Cycles | MTTR(sec) | Failure rate(%) | Auto-healing | Performance degradation |
+--------+-------------+-----------------+--------------+-------------------------+
| 1 | 97.19 | 7 | Yes | No |
+--------+-------------+-----------------+--------------+-------------------------+
| 2 | 93.87 | 6 | Yes | No |
+--------+-------------+-----------------+--------------+-------------------------+
| 3 | 92.12 | 8 | Yes | No |
+--------+-------------+-----------------+--------------+-------------------------+
| 4 | 94.51 | 6 | Yes | No |
+--------+-------------+-----------------+--------------+-------------------------+
| 5 | 98.37 | 7 | Yes | No |
+--------+-------------+-----------------+--------------+-------------------------+
**Rally report:** :download:`random_controller_freeze_keystone_150_sec.html <../../../../raw_results/reliability/rally_results/NovaServers/boot_and_delete_server/random_controller_freeze_keystone_150_sec.html>`
.. table:: **Testing results summary**
+--------+-----------+--------------+
| Value | MTTR(sec) | Failure rate |
+--------+-----------+--------------+
| Min | 92.12 | 6 |
+--------+-----------+--------------+
| Max | 98.37 | 8 |
+--------+-----------+--------------+
| SLA | Yes | No |
+--------+-----------+--------------+
**Detailed results description**
This factor affects an OpenStack cluster operations.
After the keystone processes freeze on controllers, the HA
logic needs approximately 95 seconds to recover service operation.
After recovering, performance degradation is not observed but
only at small concurrency. This behaviour is not normal
for an HA OpenStack configuration and should be investigated in future.
.. references:
.. _Rally installation documentation: https://rally.readthedocs.io/en/latest/install.html

16
doc/source/test_results/reliability/rally_plugins/scrappy.conf Normal file
View File

@ -0,0 +1,16 @@
# SSH credentials
SSH_LOGIN="root"
SSH_PASS="r00tme"
# Controller nodes
CONTROLLERS[0]="10.44.0.7"
CONTROLLERS[1]="10.44.0.6"
CONTROLLERS[2]="10.44.0.5"
# Compute nodes
COMPUTES[0]="10.44.0.3"
COMPUTES[1]="10.44.0.4"
COMPUTES[2]="10.44.0.8"
#Scrappy base path
SCRAPPY_BASE="/root/scrappy"

115
doc/source/test_results/reliability/rally_plugins/scrappy.py Normal file
View File

@ -0,0 +1,115 @@
# Copyright 2014: Mirantis Inc.
# All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License"); you may
# not use this file except in compliance with the License. You may obtain
# a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
# WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
# License for the specific language governing permissions and limitations
# under the License.
"""
Rully scrappy plugin
This is pluging was designed for OpenStack
reliability testing.
"""
from rally.common.i18n import _
from rally import consts
from rally.task import sla
import os
from rally.common import logging
from rally.common import streaming_algorithms as streaming
LOG = logging.getLogger(__name__)
class MttrCalculation():
def __init__(self):
self.min_timestamp = streaming.MinComputation()
self.max_timestamp = streaming.MaxComputation()
self.mttr = 0
self.last_error_duration = 0
self.last_iteration = None
def add(self, iteration):
if iteration["error"]:
# Store duration of last error iteration
if self.max_timestamp.result() < iteration["timestamp"]:
self.last_error_duration = iteration["duration"]
self.min_timestamp.add(iteration["timestamp"])
self.max_timestamp.add(iteration["timestamp"])
LOG.info("TIMESTAMP: %s" % iteration["timestamp"])
self.last_iteration = iteration
def result(self):
self.mttr = round(self.max_timestamp.result() -
self.min_timestamp.result() +
self.last_error_duration, 2)
# SLA Context don't have information about iterations count,
# so assume that if last iteration completed with error,
# that cluster was not auto-healed
if self.last_iteration["error"]:
self.mttr = "Inf."
return(self.mttr)
@sla.configure(name="scrappy")
class Scrappy(sla.SLA):
"""Scrappy events."""
CONFIG_SCHEMA = {
"type": "object",
"$schema": consts.JSON_SCHEMA,
"properties": {
"on_iter": {"type": "number"},
"execute": {"type": "string"},
"cycle": {"type": "number"}
}
}
def __init__(self, criterion_value):
super(Scrappy, self).__init__(criterion_value)
self.on_iter = self.criterion_value.get("on_iter", None)
self.execute = self.criterion_value.get("execute", None)
self.cycle = self.criterion_value.get("cycle", 0)
self.errors = 0
self.total = 0
self.error_rate = 0.0
self.mttr = MttrCalculation()
def add_iteration(self, iteration):
self.total += 1
if iteration["error"]:
self.errors += 1
self.mttr.add(iteration)
"""Start iteration event"""
if self.on_iter == self.total:
LOG.info("Scrappy testing cycle: ITER: %s" % self.cycle)
LOG.info("Scrappy executing: %s" % self.on_iter)
os.system(self.execute)
self.error_rate = self.errors * 100.0 / self.total
self.success = self.error_rate <= 5
return self.success
def merge(self, other):
self.total += other.total
self.errors += other.errors
if self.total:
self.error_rate = self.errors * 100.0 / self.total
self.success = self.error_rate <= 5
return self.success
def details(self):
return (_("Scrappy failure rate %.2f%% MTTR %s seconds - %s") %
(self.error_rate, self.mttr.result(), self.status()))

116
doc/source/test_results/reliability/rally_plugins/scrappy.sh Executable file
View File

@ -0,0 +1,116 @@
# Copyright 2014: Mirantis Inc.
# All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License"); you may
# not use this file except in compliance with the License. You may obtain
# a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
# WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
# License for the specific language governing permissions and limitations
# under the License.
#!/bin/bash -xe
# source credentionals
if [ -f /data/rally/rally_plugins/scrappy/scrappy.conf ];
then
. /data/rally/rally_plugins/scrappy/scrappy.conf
else
exit -1
fi
#
# Function exetute command over ssh
# Login & password stored in scrappy.conf
#
function ssh_exec() {
local ssh_node=$1
local ssh_cmd=$2
local ssh_options='-oConnectTimeout=5 -oStrictHostKeyChecking=no -oCheckHostIP=no -oUserKnownHostsFile=/dev/null -oRSAAuthentication=no'
echo "sshpass -p ${SSH_PASS} ssh ${ssh_options} ${SSH_LOGIN}@${ssh_node} ${ssh_cmd}"
local ssh_result=`sshpass -p ${SSH_PASS} ssh ${ssh_options} ${SSH_LOGIN}@${ssh_node} ${ssh_cmd}`
echo "$ssh_result"
}
#
# Function return random controller node from Fuel cluster
#
function get_random_controller() {
local random_controller=${CONTROLLERS[$RANDOM % ${#CONTROLLERS[@]}]}
echo $random_controller
}
#
# Function return random compute node from Fuel cluster
#
function get_random_compute() {
local random_compute=${COMPUTES[$RANDOM % ${#COMPUTES[@]}]}
echo $random_compute
}
#
# Factors
#
function random_controller_kill_rabbitmq() {
local action=$1
local controller_node=$(get_random_controller)
local result=`ssh_exec ${controller_node} "${SCRAPPY_BASE}/scrappy_host.sh send_signal rabbitmq_server -KILL"`
echo "$result"
}
function random_controller_freeze_process_random_interval() {
local process_name=$1
local interval=$2
local controller_node=$(get_random_controller)
local result=`ssh_exec ${controller_node} "${SCRAPPY_BASE}/scrappy_host.sh freeze_process_random_interval ${process_name} ${interval}"`
echo "$result"
}
function random_controller_freeze_process_fixed_interval() {
local process_name=$1
local interval=$2
local controller_node=$(get_random_controller)
local result=`ssh_exec ${controller_node} "${SCRAPPY_BASE}/scrappy_host.sh freeze_process_fixed_interval ${process_name} ${interval}"`
echo "$result"
}
function random_controller_reboot() {
local controller_node=$(get_random_controller)
local result=`ssh_exec ${controller_node} "${SCRAPPY_BASE}/scrappy_host.sh reboot_node"`
echo "$result"
}
function usage() {
echo "usage"
echo "TODO"
}
#
# Main
#
function main() {
local factor=$1
case ${factor} in
random_controller_kill_rabbitmq)
random_controller_kill_rabbitmq $2
;;
random_controller_freeze_process_random_interval)
random_controller_freeze_process_random_interval $2 $3
;;
random_controller_freeze_process_fixed_interval)
random_controller_freeze_process_fixed_interval $2 $3
;;
random_controller_reboot)
random_controller_reboot
;;
*)
usage
;;
esac
}
main "$@"

132
doc/source/test_results/reliability/rally_plugins/scrappy_host.sh Executable file
View File

@ -0,0 +1,132 @@
# Copyright 2014: Mirantis Inc.
# All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License"); you may
# not use this file except in compliance with the License. You may obtain
# a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
# WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
# License for the specific language governing permissions and limitations
# under the License.
#!/bin/bash -xe
LOG_FILE="/var/log/scrappy.log"
#
# Logging function
#
function log() {
echo "`date -u` scrappy_host: $1" >> ${LOG_FILE}
}
#
# This is function send specified signal
# to all processes with given name
#
function send_signal() {
local process_name=$1
local signal=$2
local pids=`ps -ef | grep $process_name | grep -v grep | grep -v scrappy_host | awk '{print $2}'`
for each_pid in ${pids};
do
log "sending signal: ${signal} to ${process_name} with pid:$each_pid"
kill ${signal} ${each_pid}
done
}
#
# This is function control services
#
function service_control() {
local service_name=$1
local service_action=$2
log "service control: $service_name action: $service_action"
service $service_name $service_action
}
#
# This is function reboot node
#
function reboot_node() {
log "reboot"
shutdown -r now
}
#
# This factor freeze specifid process
#
function freeze_process_random_interval {
local process_name=$1
local max_interval=$2
local interval=$(( ($RANDOM % ${max_interval}) + 1))
log "freeze_process_random_interval: freezing process ${process_name} freeze interval ${interval}"
send_signal ${process_name} '-STOP'
sleep ${interval}
log "freeze_process_random_interval: unfreezing process ${process_name}"
send_signal ${process_name} '-CONT'
}
#
# This factor freeze specifid process
#
function freeze_process_fixed_interval {
local process_name=$1
local interval=$2
log "freeze_process_fixed_interval: freezing process ${process_name} freeze interval ${interval}"
send_signal ${process_name} '-STOP'
sleep ${interval}
log "freeze_process_fixed_interval: unfreezing process ${process_name}"
send_signal ${process_name} '-CONT'
}
#
# Show usage
#
function usage() {
echo "scrappy_host usage:"
echo "scrappy_host commands:"
echo -e "\t send_signal process_name signal"
echo -e "\t service_control service_name action"
echo -e "\t freeze_process_random_interval process max_interval"
echo -e "\t freeze_process_fixed_interval process interval"
echo -e "\t reboot_node"
}
#
# main
#
function main() {
local command=$1
case $command in
send_signal)
send_signal $2 $3
;;
service_control)
service_control $2 $3
;;
reboot_node)
reboot_node
;;
freeze_process_random_interval)
set +xe
freeze_process_random_interval $2 $3 &
set -xe
;;
freeze_process_fixed_interval)
set +xe
freeze_process_fixed_interval $2 $3 &
set -xe
;;
*)
usage
exit -1
;;
esac
}
main "$@"

37
doc/source/test_results/reliability/rally_scenarios/NovaServers/boot_and_delete_server/random_controller_freeze_keystone_150_sec.json Normal file
View File

@ -0,0 +1,37 @@
{% set flavor_name = flavor_name or "m1.tiny" %}
{% set image_name = image_name or "^(cirros.*uec|TestVM)$" %}
{
"NovaServers.boot_and_delete_server": [
{% for i in range (0, 5, 1) %}
{
"args": {
"flavor": {
"name": "{{flavor_name}}"
},
"image": {
"name": "{{image_name}}"
},
"force_delete": false
},
"runner": {
"type": "constant",
"times": 100,
"concurrency": 5
},
"context": {
"users": {
"tenants": 1,
"users_per_tenant": 1
}
},
"sla": {
"scrappy": {
"on_iter": 20,
"execute": "/bin/bash /data/rally/rally_plugins/scrappy/scrappy.sh random_controller_freeze_process_fixed_interval keystone 150",
"cycle": {{i}}
},
}
},
{% endfor %}
]
}

37
doc/source/test_results/reliability/rally_scenarios/NovaServers/boot_and_delete_server/random_controller_freeze_memcached_150_sec.json Normal file
View File

@ -0,0 +1,37 @@
{% set flavor_name = flavor_name or "m1.tiny" %}
{% set image_name = image_name or "^(cirros.*uec|TestVM)$" %}
{
"NovaServers.boot_and_delete_server": [
{% for i in range (0, 5, 1) %}
{
"args": {
"flavor": {
"name": "{{flavor_name}}"
},
"image": {
"name": "{{image_name}}"
},
"force_delete": false
},
"runner": {
"type": "constant",
"times": 100,
"concurrency": 5
},
"context": {
"users": {
"tenants": 1,
"users_per_tenant": 1
}
},
"sla": {
"scrappy": {
"on_iter": 20,
"execute": "/bin/bash /data/rally/rally_plugins/scrappy/scrappy.sh random_controller_freeze_process_fixed_interval memcached 150",
"cycle": {{i}}
},
}
},
{% endfor %}
]
}

37
doc/source/test_results/reliability/rally_scenarios/NovaServers/boot_and_delete_server/random_controller_freeze_nova-api_150_sec.json Normal file
View File

@ -0,0 +1,37 @@
{% set flavor_name = flavor_name or "m1.tiny" %}
{% set image_name = image_name or "^(cirros.*uec|TestVM)$" %}
{
"NovaServers.boot_and_delete_server": [
{% for i in range (0, 5, 1) %}
{
"args": {
"flavor": {
"name": "{{flavor_name}}"
},
"image": {
"name": "{{image_name}}"
},
"force_delete": false
},
"runner": {
"type": "constant",
"times": 100,
"concurrency": 15
},
"context": {
"users": {
"tenants": 1,
"users_per_tenant": 1
}
},
"sla": {
"scrappy": {
"on_iter": 20,
"execute": "/bin/bash /data/rally/rally_plugins/scrappy/scrappy.sh random_controller_freeze_process_fixed_interval nova-api 150",
"cycle": {{i}}
},
}
},
{% endfor %}
]
}

38
doc/source/test_results/reliability/rally_scenarios/NovaServers/boot_and_delete_server/random_controller_kill_mysqld.json Normal file
View File

@ -0,0 +1,38 @@
{% set flavor_name = flavor_name or "m1.tiny" %}
{% set image_name = image_name or "^(cirros.*uec|TestVM)$" %}
{
"NovaServers.boot_and_delete_server": [
{% for i in range (0, 5, 1) %}
{
"args": {
"flavor": {
"name": "{{flavor_name}}"
},
"image": {
"name": "{{image_name}}"
},
"force_delete": false
},
"runner": {
"type": "constant",
"times": 100,
"concurrency": 5
},
"context": {
"users": {
"tenants": 1,
"users_per_tenant": 1
}
},
"sla": {
"scrappy": {
"on_iter": 20,
"execute": "/bin/bash /data/rally/rally_plugins/scrappy/scrappy.sh send_signal /usr/sbin/mysqld -KILL",
"cycle": {{i}}
}
}
},
{% endfor %}
]
}

38
doc/source/test_results/reliability/rally_scenarios/NovaServers/boot_and_delete_server/random_controller_kill_rabbitmq.json Normal file
View File

@ -0,0 +1,38 @@
{% set flavor_name = flavor_name or "m1.tiny" %}
{% set image_name = image_name or "^(cirros.*uec|TestVM)$" %}
{
"NovaServers.boot_and_delete_server": [
{% for i in range (0, 5, 1) %}
{
"args": {
"flavor": {
"name": "{{flavor_name}}"
},
"image": {
"name": "{{image_name}}"
},
"force_delete": false
},
"runner": {
"type": "constant",
"times": 100,
"concurrency": 5
},
"context": {
"users": {
"tenants": 1,
"users_per_tenant": 1
}
},
"sla": {
"scrappy": {
"on_iter": 20,
"execute": "/bin/bash /data/rally/rally_plugins/scrappy/scrappy.sh random_controller_kill_rabbitmq",
"cycle": {{i}}
}
}
},
{% endfor %}
]
}

38
doc/source/test_results/reliability/rally_scenarios/NovaServers/boot_and_delete_server/random_controller_reboot_factor.json Normal file
View File

@ -0,0 +1,38 @@
{% set flavor_name = flavor_name or "m1.tiny" %}
{% set image_name = image_name or "^(cirros.*uec|TestVM)$" %}
{
"NovaServers.boot_and_delete_server": [
{% for i in range (0, 5, 1) %}
{
"args": {
"flavor": {
"name": "{{flavor_name}}"
},
"image": {
"name": "{{image_name}}"
},
"force_delete": false
},
"runner": {
"type": "constant",
"times": 100,
"concurrency": 5
},
"context": {
"users": {
"tenants": 1,
"users_per_tenant": 1
}
},
"sla": {
"scrappy": {
"on_iter": 20,
"execute": "/bin/bash /data/rally/rally_plugins/scrappy/scrappy.sh random_controller_reboot",
"cycle": {{i}}
}
}
},
{% endfor %}
]
}

856
raw_results/reliability/rally_results/NovaServers/boot_and_delete_server/random_controller_freeze_keystone_150_sec.html Normal file
View File

File diff suppressed because one or more lines are too long

856
raw_results/reliability/rally_results/NovaServers/boot_and_delete_server/random_controller_freeze_memcached_150_sec.html Normal file
View File

File diff suppressed because one or more lines are too long

856
raw_results/reliability/rally_results/NovaServers/boot_and_delete_server/random_controller_freeze_nova_api_150_sec.html Normal file
View File

File diff suppressed because one or more lines are too long

856
raw_results/reliability/rally_results/NovaServers/boot_and_delete_server/random_controller_kill_mysqld.html Normal file
View File

File diff suppressed because one or more lines are too long

856
raw_results/reliability/rally_results/NovaServers/boot_and_delete_server/random_controller_kill_rabbitmq.html Normal file
View File

File diff suppressed because one or more lines are too long

856
raw_results/reliability/rally_results/NovaServers/boot_and_delete_server/reboot_random_controller.html Normal file
View File

File diff suppressed because one or more lines are too long