-
Notifications
You must be signed in to change notification settings - Fork 0
/
create_volume_backup.yaml
230 lines (214 loc) · 10.3 KB
/
create_volume_backup.yaml
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
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
---
version: "2.0"
create_volume_backup:
type: direct
description: |
Version 2.3
Created by EveryWare AG, 2019-09-27, [email protected]
2020-01-14 add rotate parameter
2020-09-02 return task(success_task).published.backup_status
2020-09-04 chg sender to [email protected]
2021-02-04 improve error handling
2021-03-12 use cinder.backups_delete in an sequential loop instead of async
This workflow creates a backup of a volume to the swift service.
It actually creates a temporary snapshot of a volume and then moves it to swift.
References:
* https://docs.openstack.org/mistral/latest/user/wf_lang_v2.html
* https://yaql.readthedocs.io/en/latest/language_reference.html
* https://gitlab.cern.ch/cloud-infrastructure/mistral-workflows/blob/master/workflows/instance_snapshot.yaml
* https://docs.stackstorm.com/workflows.html
Description of the input params:
* volume_id: volume to backup
* name: name of volume backup ($pattern will be appended)
* incremental: number of incremental backups between full backups (0 for full only)
* rotate: number of backups to keep (0 to disable rotation);
you should respect: rotate % (incremental + 1) = 0
if incremental > 0 then nbr_backup is in an interval instead of constant:
=> nbr_backup: [rotate, rotate + incremental]
=> nbr_full: [rotate % incremental, rotate % incremental + 1]
=> nbr_incremental: [rotate - rotate % incremental, rotate + incremental - rotate % incremental - 1]
* force: true if backup allow when the volume is attached to a running VM
* to_addr_success: email address of success or null
* to_addr_error: email address of error or null
input:
- volume_id
- name
- incremental: 27
- force: true
- rotate: 28
- to_addr_success: null
- to_addr_error: null
- pattern: "{0}_backup_{1}"
output:
#everything: <% $ %>
backup_id: <% task(success_task).published.backup_id %>
backup_status: <% task(success_task).published.backup_status %>
output-on-error:
backup_id: <% task(error_task).published.backup_id %>
backup_status: <% task(error_task).published.backup_status %>
tasks:
init:
description: 'Validate and convert input paramters. Define variables.'
action: std.noop
publish:
incremental: <% int($.incremental) %>
force: <% let(var => $.force) -> switch(isBoolean($var) => $var, not isBoolean($var) => $var.toLower() = 'true') %>
rotate: <% int($.rotate) %>
backup_name: <% $.pattern.format($.name, now().format("%Y%m%d%H%M")) %>
backup_regex: <% $.pattern.format($.name, '*') %>
on-success: backup_list
backup_list:
description: 'Get list of volume backups matching backup_regex.'
action: cinder.backups_list
publish:
backup_list: <% task().result.where($.status='available').where(regex(task(init).published.backup_regex).matches($.name)).orderBy($.created_at).id %>
on-success:
- rotate_backup: <% $.rotate >= 1 %>
- create_backup: <% $.rotate = 0 %>
on-error: error_task
# nok: all backups_delete are run async and all but the first fail because of incremental chain dependencies
# rotate_backup_old:
# description: 'Rotate the volume backups.'
# with-items: backup_id2 in <% $.backup_list.take( [$.backup_list.len() + 1 - $.rotate, 0].max() ).reverse() %>
# concurrency: 1
# wait-before: 10
# action: cinder.backups_delete
# input:
# backup: <% $.backup_id2 %>
# force: false
# on-complete: wait_on_delete_old
#
# wait_on_delete_old:
# description: |
# Because cinder.backups_delete is async, it "succeeds" immediately.
# Therefore we need to poll here the backup_status via cinder.backups_list()
# and continue waiting as long as there are some backups with status "deleting" and timeout is not reached.
# action: cinder.backups_list
# retry:
# delay: 60
# count: <% int(60*1) %>
# continue-on: <% task().result.where($.status='deleting').where(regex(task(init).published.backup_regex).matches($.name)).len() >= 1 %>
# on-complete: create_backup
rotate_backup:
description: |
Rotate the volume backups.
backup_list is sorted by created_at; we want to keep only the last 'rotate - 1' elements,
so that after the creation a a new backup we have exactly 'rotate' backups.
Therefore we need to delete the first 'backup_list.len + 1 - rotate' elements.
We reverse the resulted list to deal with incremental dependencies.
action: std.noop
publish:
delbackup_list: <% $.backup_list.take( [$.backup_list.len() + 1 - $.rotate, 0].max() ).reverse() %>
on-success:
- delete_backup: <% $.delbackup_list.len() >= 1 %>
- create_backup: <% $.delbackup_list.len() = 0 %>
delete_backup:
description: |
Delete first element of a volume backup list, recursively, until the list is empty.
Ignore delete errors.
action: cinder.backups_delete
input:
backup: <% $.delbackup_list.first() %>
force: false
publish:
delbackup_list: <% $.delbackup_list.skip(1) %>
publish-on-error:
delbackup_list: <% $.delbackup_list.skip(1) %>
on-complete:
- wait_on_delete
wait_on_delete:
description: |
Because cinder.backups_delete is async, it "succeeds" immediately.
Therefore we need to poll here the backup_status via cinder.backups_list()
and continue waiting as long as there are some backups with status "deleting" and timeout is not reached.
action: cinder.backups_list
retry:
delay: 60
count: <% int(60*1) %>
continue-on: <% task().result.where($.status='deleting').where(regex(task(init).published.backup_regex).matches($.name)).len() >= 1 %>
publish:
backup_list: <% task().result.where($.status='available').where(regex(task(init).published.backup_regex).matches($.name)).orderBy($.created_at).id %>
publish-on-error:
backup_id: null
backup_status: <% task().result %>
on-success:
- delete_backup: <% $.delbackup_list.len() >= 1 %>
- create_backup: <% $.delbackup_list.len() = 0 %>
on-error: error_task
create_backup:
description: 'Creates a volume backup.'
action: cinder.backups_create
input:
volume_id: <% $.volume_id %>
name: <% $.backup_name %>
incremental: <% $.backup_list.len() mod ($.incremental + 1) != 0 %>
force: <% $.force %>
publish:
backup_id: <% task().result.id %>
backup_status: <% task().result.status %>
publish-on-error:
backup_id: null
backup_status: <% task().result %>
on-success: wait_on_backup
on-error: error_task
wait_on_backup:
description: |
Because cinder.backups_create is async, it "succeeds" immediately.
Therefore we need to poll here the backup_status via cinder.backups_get(backup_id)
and continue waiting as long as it is "creating" and timeout is not reached.
action: cinder.backups_get
input:
backup_id: <% $.backup_id %>
retry:
delay: 60
count: <% int(60*23) %>
break-on: <% task().result.status = 'error' %>
continue-on: <% task().result.status = 'creating' %>
publish:
backup_status: <% task().result.status %>
on-success:
- success_task: <% $.backup_status = 'available' %>
- error_task: <% $.backup_status != 'available' %>
on-error: error_task
success_task:
description: 'Checks if a success mail has to be sent to the end user'
action: std.noop
publish:
backup_id: <% $.backup_id %>
backup_status: <% $.backup_status %>
on-success:
- success_mail: <% $.to_addr_success %>
- succeed
error_task:
description: 'Checks if an error mail has to be sent to the end user'
action: std.noop
publish:
backup_id: <% $.backup_id %>
backup_status: <% $.backup_status %>
on-success:
- error_mail: <% $.to_addr_error %>
- fail
success_mail:
description: 'Sends a success mail to the end user'
action: std.email
input:
to_addrs: <% $.to_addr_success.split(';') %>
subject: Volume Backup created successfully
body: |
The volume backup "<% $.backup_name %>" has been created successfully.
You can check the execution logs of the execution <% execution().id %>
-- Thanks, Openstack Team.
from_addr: '[email protected]'
smtp_server: 'smtp'
error_mail:
description: 'Sends an error mail to the end user'
action: std.email
input:
to_addrs: <% $.to_addr_error.split(';') %>
subject: Volume Backup has failed
body: |
The volume backup "<% $.backup_name %>" has failed.
Please check in the execution logs of the execution <% execution().id %>
-- Thanks, Openstack Team.
from_addr: '[email protected]'
smtp_server: 'smtp'