1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
|
.. This work is licensed under a Creative Commons Attribution 4.0 International
.. License.
.. http://creativecommons.org/licenses/by/4.0
.. (c) OPNFV, Huawei Technologies Co.,Ltd and others.
*************************************
Yardstick Test Case Description TC074
*************************************
.. _Storperf: https://wiki.opnfv.org/display/storperf/Storperf
+-----------------------------------------------------------------------------+
|Storperf |
| |
+--------------+--------------------------------------------------------------+
|test case id | OPNFV_YARDSTICK_TC074_Storperf |
| | |
+--------------+--------------------------------------------------------------+
|metric | Storage performance |
| | |
+--------------+--------------------------------------------------------------+
|test purpose | To evaluate and report on the Cinder volume performance. |
| | |
| | This testcase integrates with OPNFV StorPerf to measure |
| | block performance of the underlying Cinder drivers. Many |
| | options are supported, and even the root disk (Glance |
| | ephemeral storage can be profiled. |
| | |
| | The fundamental concept of the test case is to first fill |
| | the volumes with random data to ensure reported metrics |
| | are indicative of continued usage and not skewed by |
| | transitional performance while the underlying storage |
| | driver allocates blocks. |
| | The metrics for filling the volumes with random data |
| | are not reported in the final results. The test also |
| | ensures the volumes are performing at a consistent level |
| | of performance by measuring metrics every minute, and |
| | comparing the trend of the metrics over the run. By |
| | evaluating the min and max values, as well as the slope of |
| | the trend, it can make the determination that the metrics |
| | are stable, and not fluctuating beyond industry standard |
| | norms. |
| | |
+--------------+--------------------------------------------------------------+
|configuration | file: opnfv_yardstick_tc074.yaml |
| | |
| | * agent_count: 1 - the number of VMs to be created |
| | * agent_image: "Ubuntu-14.04" - image used for creating VMs |
| | * public_network: "ext-net" - name of public network |
| | * volume_size: 2 - cinder volume size |
| | * block_sizes: "4096" - data block size |
| | * queue_depths: "4" - the number of simultaneous I/Os |
| | to perform at all times |
| | * StorPerf_ip: "192.168.200.2" |
| | * query_interval: 10 - state query interval |
| | * timeout: 600 - maximum allowed job time |
| | |
+--------------+--------------------------------------------------------------+
|test tool | Storperf_ |
| | |
| | StorPerf is a tool to measure block and object storage |
| | performance in an NFVI. |
| | |
| | StorPerf is delivered as a Docker container from |
| | https://hub.docker.com/r/opnfv/storperf-master/tags/. |
| | |
| | The underlying tool used is FIO, and StorPerf supports |
| | any FIO option in order to tailor the test to the exact |
| | workload needed. |
| | |
+--------------+--------------------------------------------------------------+
|references | Storperf_ |
| | |
| | ETSI-NFV-TST001 |
| | |
+--------------+--------------------------------------------------------------+
|applicability | Test can be configured with different: |
| | |
| | * agent_count |
| | * volume_size |
| | * block_sizes |
| | * queue_depths |
| | * query_interval |
| | * timeout |
| | * target=[device or path] |
| | The path to either an attached storage device |
| | (/dev/vdb, etc) or a directory path (/opt/storperf) that |
| | will be used to execute the performance test. In the case |
| | of a device, the entire device will be used. If not |
| | specified, the current directory will be used. |
| | * workload=[workload module] |
| | If not specified, the default is to run all workloads. The |
| | workload types are: |
| | - rs: 100% Read, sequential data |
| | - ws: 100% Write, sequential data |
| | - rr: 100% Read, random access |
| | - wr: 100% Write, random access |
| | - rw: 70% Read / 30% write, random access |
| | measurements. |
| | * workloads={json maps} |
| | This parameter supercedes the workload and calls the V2.0 |
| | API in StorPerf. It allows for greater control of the |
| | parameters to be passed to FIO. For example, running a |
| | random read/write with a mix of 90% read and 10% write |
| | would be expressed as follows: |
| | {"9010randrw": {"rw":"randrw","rwmixread": "90"}} |
| | Note: This must be passed in as a string, so don't forget |
| | to escape or otherwise properly deal with the quotes. |
| | |
| | * report= [job_id] |
| | Query the status of the supplied job_id and report on |
| | metrics. If a workload is supplied, will report on only |
| | that subset. |
| | * availability_zone: Specify the availability zone which |
| | the stack will use to create instances. |
| | * volume_type: |
| | Cinder volumes can have different types, for example |
| | encrypted vs. not encrypted. |
| | To be able to profile the difference between the two. |
| | * subnet_CIDR: Specify subnet CIDR of private network |
| | * stack_name: Specify the name of the stack that will be |
| | created, the default: "StorperfAgentGroup" |
| | * volume_count: Specify the number of volumes per |
| | virtual machines |
| | |
| | There are default values for each above-mentioned option. |
| | |
+--------------+--------------------------------------------------------------+
|pre-test | If you do not have an Ubuntu 14.04 image in Glance, you will |
|conditions | need to add one. |
| | |
| | Storperf is required to be installed in the environment. |
| | There are two possible methods for Storperf installation: |
| | Run container on Jump Host |
| | Run container in a VM |
| | |
| | Running StorPerf on Jump Host |
| | Requirements: |
| | - Docker must be installed |
| | - Jump Host must have access to the OpenStack Controller |
| | API |
| | - Jump Host must have internet connectivity for |
| | downloading docker image |
| | - Enough floating IPs must be available to match your |
| | agent count |
| | |
| | Running StorPerf in a VM |
| | Requirements: |
| | - VM has docker installed |
| | - VM has OpenStack Controller credentials and can |
| | communicate with the Controller API |
| | - VM has internet connectivity for downloading the |
| | docker image |
| | - Enough floating IPs must be available to match your |
| | agent count |
| | |
| | No POD specific requirements have been identified. |
| | |
+--------------+--------------------------------------------------------------+
|test sequence | description and expected result |
| | |
+--------------+--------------------------------------------------------------+
|step 1 | Yardstick calls StorPerf to create the heat stack with the |
| | number of VMs and size of Cinder volumes specified. The |
| | VMs will be on their own private subnet, and take floating |
| | IP addresses from the specified public network. |
| | |
+--------------+--------------------------------------------------------------+
|step 2 | Yardstick calls StorPerf to fill all the volumes with |
| | random data. |
| | |
+--------------+--------------------------------------------------------------+
|step 3 | Yardstick calls StorPerf to perform the series of tests |
| | specified by the workload, queue depths and block sizes. |
| | |
+--------------+--------------------------------------------------------------+
|step 4 | Yardstick calls StorPerf to delete the stack it created. |
| | |
+--------------+--------------------------------------------------------------+
|test verdict | None. Storage performance results are fetched and stored. |
| | |
+--------------+--------------------------------------------------------------+
|
