1 Use Cases and Scenarios
2 -----------------------
4 This section describes the use cases and scenarios to verify the
5 requirements of Escalator.
9 1. Upgrade a system with HA configuration
10 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
12 A HA configuration system is very popular in the operator's data centre.
13 It is a typical product environment. It is always running 7\*24 with VNFs
14 running on it to provide services to the end users.
17 2. Upgrade a system with non-HA configuration
18 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
20 A non-HA configuration system is normally deployed for experimental or
21 development usages, such as a Vagrant/VM environment.
23 Escalator supports the upgrade in this scenario, but it does not guarantee a
28 Use case #1: Smooth upgrade in a HA configuration
29 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
30 For a system with HA configuration, the operator can use Escalator to
31 smooth-upgrade NFVI/VIM components into a new version without any service
34 When a compute node being upgraded, the VMs on the node may need to be migrated
35 to other compute nodes to avoid service outage, so it is requred that there are
36 enough redundant resources to migrate VMs on this compute node.
38 Before upgrade, the operator can use Escalator to check whether smooth upgrade
39 conditions are all satisfied. These conditions include whether there are enough
40 idle resources to migrate VMs during updrading, and whether the new version is
41 compatible with the current one, etc. If there are some conditions not
42 satisfied, Escalator will show them. Escalator can also provide the solutions if
43 there is any, such as the number and configuration of spare compute nodes which
46 When upgrade starts, Escalator will also automatically check whether smooth
47 upgrade conditions are all satisfied. If some smooth upgrade conditions are not
48 satisfied, Escalator will show the failure of smooth upgrade.
52 1. The system is running as normal.
53 2. The VNFs are providing services as usual.
57 1. The VNFs are continually providing services during the upgrade.
58 2. The operator successfully logged in the GUI of Escalator to select the
59 software packages including Linux OS, Hypervisor, OpenStack, ODL and other
60 OPNFV components, ect. (All or part of components could be selected.)
61 3. Select the nodes to be upgraded. i.e. controller node, network node,
62 storage node and compute node, etc.
63 4. Select "Disable Scale-up". It will limit the scale-up operation when
64 upgrade is in progress to prevent failures due to the shortage of
66 5. Select "Check Smooth Upgrade Conditions". If Escalator shows that there are
67 some conditions not satisfied, try to resolve them according to the
69 6. Select "Smooth Upgrade", then apply the upgrade operation.
70 7. Select "Restore Scale-up" after the upgrade. It will restore scale-up to
71 the original enabled/disabled state before upgrade.
75 1. The system is upgraded successfully.
76 2. There is no service outage during the upgrade.
77 3. The VNFs are providing services as usual after the upgrade.
79 Use case #2: Roll-back after a failed smooth upgrade in a HA configuration
80 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
81 For a system with HA configuration, if the upgrade fails when the operator is
82 smooth-upgrading NFVI/VIM components into a new version using Escalator, the
83 operator can roll-back the system without any service outage.
87 1. The system is running as normal.
88 2. The VNFs are providing services as usual.
89 3. Scale-up operation is disabled.
90 4. Smooth upgrade failed.
94 1. Escalator concludes that the upgrade has failed and provides the operator
96 2. Select the "Roll-back" operation.
97 3. If the roll-back is successful, go to step 4, otherwise the operator can
98 select "Restore Backup" to restore the system from the backup data.
99 4. Select "Restore Scale-up" after the roll-back. It will restore scale-up to
100 the original enabled/disabled state before upgrade.
104 1. The system is rolled-back successfully when the upgrade failed.
105 2. There is no service outage during the roll-back.
106 3. The VNFs are providing services as usual after the roll-back.
108 Use case #3: Roll-back after a successful smooth upgrade in a HA configuration
109 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
110 When a smooth upgrade in a HA configuration is successful, the operator may want
111 to roll-back for some reasons, such as performance issues.
112 Escalator supports roll-back after a successful smooth upgrade without any
117 1. The system is running as normal.
118 2. The VNFs are providing services as usual.
119 3. Smooth upgrade succeeded.
123 1. Select "Disable Scale-up". It will limit the scale-up operation when roll-
124 back is in progress to prevent failures due to the shortage of resources.
125 2. Select "Check Smooth Roll-back Conditions". If Escalator shows that there
126 are some conditions not satisfied, try to resolve them according to the
128 3. Select "Roll-back", then apply the roll-back operation.
129 4. If the roll-back is successful, go to step 5, otherwise the operator can
130 select "Restore Backup" to restore the system from the backup data.
131 5. Select "Restore Scale-up" after the roll-back. It will restore scale-up to
132 the original enabled/disabled state before roll-back.
136 1. The system is rolled-back successfully.
137 2. There is no service outage during the roll-back.
138 3. The VNFs are providing services as usual after the roll-back.
140 Use case #4: Non-smooth upgrade in a non-HA/HA configuration
141 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
142 For a system with non-HA configuration, the operator can also use Escalator to
143 upgrade NFVI/VIM components into a new version. In this case, the upgrade may
144 result in service outage. In other words, the upgrade is non-smooth.
145 For a system with HA configuration, if the service outage is acceptable or
146 inevitable, the operator can also use Escalator to non-smoothly upgrade the
151 1. The system is running as normal.
155 1. The operator successfully logged in the GUI of Escalator to select the
156 software packages including Linux OS, Hypervisor, OpenStack, ODL and other
157 OPNFV components, ect. (All or part of components could be selected.)
158 2. Select the nodes to be upgraded. i.e. controller node, network node,
159 storage node and compute node, etc.
160 3. Select "Non-Smooth Upgrade", then apply the upgrade operation.
164 1. The system is upgraded successfully.
166 Use case #5: Roll-back after a failed non-smooth upgrade in a non-HA/HA configuration
167 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
168 For a system with non-HA/HA configuration, if the upgrade fails when the
169 operator is non-smoothly upgrading NFVI/VIM components into a new version using
170 Escalator, the operator can roll-back the system. In this case, the roll-back
171 may result in service outage.
175 1. The system is running as normal.
176 2. Non-smooth upgrade failed.
180 1. Escalator concludes that the upgrade has failed and provides the operator
182 2. Select the "Roll-back" operation.
183 3. If the roll-back fails, the operator can select "Restore Backup" to restore
184 the system from the backup data.
188 1. The system is rolled-back successfully when the upgrade failed.
190 Use case #6: Roll-back after a successful non-smooth upgrade in a non-HA/HA configuration
191 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
192 When a non-smooth upgrade in a non-HA/HA configuration is successful, the
193 operator may want to roll-back for some reasons, such as performance issues.
194 Escalator supports roll-back after a successful non-smooth upgrade. In this
195 case,the roll-back may result in service outage.
199 1. The system is running as normal.
200 2. Non-smooth upgrade succeeded.
204 1. Select the "Roll-back" operation.
205 2. If the roll-back fails, the operator can select "Restore Backup" to restore
206 the system from the backup data.
210 1. The system is rolled-back successfully when the upgrade failed.