深入GitLabCI-談描述檔重構與流水線加速

本次的講題是在 「台中自由軟體愛好者社群」所做的分享，內容主要著重在 GitLab CI 在流水線加速與描述檔重構兩個部分。

使用之 .gitlab-ci.yml 範例連結：https://gitlab.com/mouson-gitlab-playground/gitlab-ci-demo-sample/-/tree/refactor-gitlab-ci-sample-v3

影片紀錄連結：https://www.youtube.com/watch?v=8wLGw7oYJKg&feature=youtu.be&t=2438

Transcript

1. 深入 GitLab CI 談描述檔重構與流⽔線加速 墨嗓 (陳佑⽵) 2020/11/28 @ 台中⾃由軟體愛好者社群
2. About Me • 陳佑⽵，朋友都叫我墨嗓(Mouson)。 專注於資訊科技整合、應⽤、開發⽅ 法及研究的T型⼈。 妞妞
3. 歡迎加入 GitLab Hero
4. . 社群 . https://fb.com/groups/GitLabTaipei https://t.me/GitLabTaiwan
5. 進入主題
6. 深入 GitLab CI • Part I: 流⽔線 (Pipeline) 加速 • Part II: .gitlab-ci.yml 描述檔如何重構
7. 怎麼讓流⽔線 (Pipeline) 更快⼀點？ Part I
8. 流⽔線加速三要素
9. 流⽔線加速三要素 • ⼯作 少做點
10. 流⽔線加速三要素 • ⼯作 少做點 • 事情 作快點
11. 流⽔線加速三要素 • ⼯作 少做點 • 事情 作快點 • 作事的 多⼀點
12. 流⽔線 Pipeline 啟動到結束 哪些地⽅花了時間？
13. 拆解
14. .gitlab-ci.yml
15. .gitlab-ci.yml
16. .gitlab-ci.yml
17. .gitlab-ci.yml Job Job Job Job Job Job Job Queue
18. .gitlab-ci.yml Job Job Job Job Job Job Job Queue Runner Server Runner Server Runners Runner Runner Runner Runner
19. .gitlab-ci.yml Job Job Job Job Job Job Job Queue Runner Server Runner Server Runners Runner Runner Runner Runner Run Job ____________________________
20. .gitlab-ci.yml Job Job Job Job Job Job Job Queue
21. •什麼⼯作先作比較好？ .gitlab-ci.yml Job Job Job Job Job Job Job Queue
22. 什麼⼯作先做比較好？
23. 什麼⼯作先做比較好？ • 跑的快的⼯作
24. 什麼⼯作先做比較好？ • 跑的快的⼯作 • 早點做完早點確認結果
25. 什麼⼯作先做比較好？ • 跑的快的⼯作 • 早點做完早點確認結果 • 有決定性的⼯作
26. 什麼⼯作先做比較好？ • 跑的快的⼯作 • 早點做完早點確認結果 • 有決定性的⼯作 • 沒通過直接 Reject
27. •什麼⼯作先做比較好？ .gitlab-ci.yml Job Job Job Job Job Job Job Queue
28. •什麼⼯作先做比較好？ •每個⼯作都是該做的嗎？ .gitlab-ci.yml Job Job Job Job Job Job Job Queue
29. 每個⼯作都是該做的嗎？
30. 每個⼯作都是該做的嗎？ • 同⼀專案中包含前端及後端的測試
31. 每個⼯作都是該做的嗎？ • 同⼀專案中包含前端及後端的測試 • 可以考慮對應原始碼變更才執⾏對應的測試
32. 每個⼯作都是該做的嗎？ • 同⼀專案中包含前端及後端的測試 • 可以考慮對應原始碼變更才執⾏對應的測試
33. •什麼⼯作先做比較好？ •每個⼯作都是該做的嗎？ .gitlab-ci.yml Job Job Job Job Job Job Job Queue
34. •什麼⼯作先做比較好？ •每個⼯作都是該做的嗎？ •有沒有⼯作花太久時間等 待進 Queue ? .gitlab-ci.yml Job Job Job Job Job Job Job Queue
35. ⼯作花太久時間等待進 Queue?
36. ⼯作花太久時間等待進 Queue? build test deploy build_a test_a deploy_a build_b test_b deploy_b build_c test_c deploy_c
37. ⼯作花太久時間等待進 Queue? build test deploy build_a test_a deploy_a build_b test_b deploy_b build_c test_c deploy_c build_b
38. ⼯作花太久時間等待進 Queue? build test deploy build_a test_a deploy_a build_b test_b deploy_b build_c test_c deploy_c build_b
39. ⼯作花太久時間等待進 Queue? • 套⽤ DAG 機制 (GitLab 12.2） build test deploy build_a test_a deploy_a build_b test_b deploy_b build_c test_c deploy_c build_b
40. ⼯作花太久時間等待進 Queue? • 套⽤ DAG 機制 (GitLab 12.2） build test deploy build_a test_a deploy_a build_b test_b deploy_b build_c test_c deploy_c build_b
41. ⼯作花太久時間等待進 Queue? • 套⽤ DAG 機制 (GitLab 12.2） build test deploy build_a test_a deploy_a build_b test_b deploy_b build_c test_c deploy_c build_b
42. .gitlab-ci.yml Job Job Job Job Job Job Job Queue
43. .gitlab-ci.yml Job Job Job Job Job Job Job Queue Runner Server Runner Server Runners Runner Runner Runner Runner
44. Job Job Job Job Job Job Job Queue Runner Server Runner Server Runners Runner Runner Runner Runner
45. Job Job Job Job Job Job Job Queue Runner Server Runner Server Runners Runner Runner Runner Runner •Runner 數量夠嗎？
46. Job Job Job Job Job Job Job Queue Runner Server Runner Server Runners Runner Runner Runner Runner •Runner 數量夠嗎？ •能執⾏的 Runner 夠嗎？
47. Job Job Job Job Job Job Job Queue Runner Server Runner Server Runners Runner Runner Runner Runner
48. Job Job Job Job Job Job Job Queue Runner Server Runner Server Runners Runner Runner Runner Runner Run Job ____________________________
49. Runner Server Runner Server Runners Runner Runner Runner Runner Run Job ____________________________
50. Runner Server Runner Server Runners Runner Runner Runner Runner Run R u n ____________________________
51. Runner Server Runner Server Runners Runner Runner Runner Runner Run R u n ____________________________ * prepare p r e p a r e p r e
52. Runner Server Runner Server Runners Runner Runner Runner Runner Run R u n ____________________________ * prepare p r e p a r e p r e * git g i t g i t r
53. Runner Server Runner Server Runners Runner Runner Runner Runner Run R u n ____________________________ * prepare p r e p a r e p r e * git g i t g i t r * restore r e s t o
54. Runner Server Runner Server Runners Runner Runner Runner Runner Run R u n ____________________________ * prepare p r e p a r e p r e * git g i t g i t r * restore r e s t o * download d o w n l o a d
55. Runner Server Runner Server Runners Runner Runner Runner Runner Run R u n ____________________________ * prepare p r e p a r e p r e * git g i t g i t r * restore r e s t o * download d o w n l o a d * script s c r i p t
56. Runner Server Runner Server Runners Runner Runner Runner Runner Run R u n ____________________________ * prepare p r e p a r e p r e * git g i t g i t r * restore r e s t o * download d o w n l o a d * script s c r i p t * package p a c k a g e
57. Runner Server Runner Server Runners Runner Runner Runner Runner Run R u n ____________________________ * prepare p r e p a r e p r e * git g i t g i t r * restore r e s t o * download d o w n l o a d * script s c r i p t * package p a c k a g e * saving s a v i n
58. Runner Server Runner Server Runners Runner Runner Runner Runner Run R u n ____________________________ * prepare p r e p a r e p r e * git g i t g i t r * restore r e s t o * download d o w n l o a d * script s c r i p t * package p a c k a g e * saving s a v i n * uploading u p l o a d i n g
59. Runner Server Runner Server Runners Runner Runner Runner Runner Run Job ____________________________ * prepare execute env * git repo fetch * restore cache * download artifacts * script execute * package install * saving cache * uploading artifacts * cleaning up f i le base var
60. •Network Performance ? Runner Server Runner Server Runners Runner Runner Runner Runner Run Job ____________________________ * prepare execute env * git repo fetch * restore cache * download artifacts * script execute * package install * saving cache * uploading artifacts * cleaning up f i le base var
61. •Network Performance ? •Docker image size ? Runner Server Runner Server Runners Runner Runner Runner Runner Run Job ____________________________ * prepare execute env * git repo fetch * restore cache * download artifacts * script execut e * package install * saving cache * uploading artifacts * cleaning up f i le base var
62. Docker Image Size ?
63. Docker Image Size ?
64. •Network Performance ? •Docker image size ? •Git Repo Size ? Runner Server Runner Server Runners Runner Runner Runner Runner Run Job ____________________________ * prepare execute env * git repo fetch * restore cache * download artifacts * script execut e * package install * saving cache * uploading artifacts * cleaning up f i le base var
65. •Network Performance ? •Docker image size ? •Git Repo Size ? •Programming Language Dependence Package Size ? Runner Server Runner Server Runners Runner Runner Runner Runner Run Job ____________________________ * prepare execute env * git repo fetch * restore cache * download artifacts * script execut e * package install * saving cache * uploading artifacts * cleaning up f i le base var
66. •Network Performance ? •Docker image size ? •Git Repo Size ? •Programming Language Dependence Package Size ? •Artifacts Size ? Runner Server Runner Server Runners Runner Runner Runner Runner Run Job ____________________________ * prepare execute env * git repo fetch * restore cache * download artifacts * script execut e * package install * saving cache * uploading artifacts * cleaning up f i le base var
67. •Network Performance ? •Docker image size ? •Git Repo Size ? •Programming Language Dependence Package Size ? •Artifacts Size ? •Script Performance ? Runner Server Runner Server Runners Runner Runner Runner Runner Run Job ____________________________ * prepare execute env * git repo fetch * restore cache * download artifacts * script execut e * package install * saving cache * uploading artifacts * cleaning up f i le base var
68. Runner Server Runner Server Runners Runner Runner Runner Runner Run Job ____________________________ * prepare execute env * git repo fetch * restore cache * download artifacts * script execut e * package install * saving cache * uploading artifacts * cleaning up f i le base var
69. .gitlab-ci.yml Job Job Job Job Job Job Job Queue Runner Server Runner Server Runners Runner Runner Runner Runner Run Job ____________________________ * prepare execute env * git repo fetch * restore cache * download artifacts * script execut e * package install * saving cache * uploading artifacts * cleaning up f i le base var
70. Part I Q&A
71. .gitlab-ci.yml 描述檔 如何重構？ Part II
72. 從前從前.... 使⽤ GitLab CI ⾃動檢查格式
73. 加入⾃動測試
74. 加入⾃動部署
75. 再增加 ⼀種版本...
76. 越來越難以維護的 .gitlab-ci.yml 該怎麼整理？
77. 怎麼整理？ 針對 v73, v74 分別有： lint、test、deploy 三種類型的⼯作
78. 1. 擷取⼀致的 部分到 Default: 擷取⼀致的 script 變為 default:
79. 1. 擷取⼀致的 部分到 Default: 擷取⼀致的 script 變為 default:
80. Default: 1. Pipeline 中所有⼯作都會套⽤ 2. 如不符使⽤，可以覆寫 3. ⽀援：image、services、 before_script、after_script、tags、 cache、artifacts、retry、timeout、 interruptible
81. 2. 整理 script 共同之程序 整理出來的 Script，可以做程序的分 組，⽅便後續使⽤
82. 2. 整理 script 共同之程序 整理出來的 Script，可以做程序的分 組，⽅便後續使⽤
83. 2. 整理 script 共同之程序 整理出來的 Script，可以做程序的分 組，⽅便後續使⽤
84. 3. 擷取重複段 落以繼承再利⽤ 利⽤ hidden job 整理 job 共同的部分 並使⽤ extends 繼承
85. 3. 擷取重複段 落以繼承再利⽤ 利⽤ hidden job 整理 job 共同的部分 並使⽤ extends 繼承
86. 3. 擷取重複段 落以繼承再利⽤ 利⽤ hidden job 整理 job 共同的部分 並使⽤ extends 繼承
87. hidden job 1. 以 「.」開頭的⼯作名稱： 整個⼯作都不執⾏
88. extends:
89. extends: 1. 把繼承來源的內容抄⼀次
90. extends: 1. 把繼承來源的內容抄⼀次 2. 可以允許既成多個來源
91. extends: 1. 把繼承來源的內容抄⼀次 2. 可以允許既成多個來源 3. 重複的內容：
92. extends: 1. 把繼承來源的內容抄⼀次 2. 可以允許既成多個來源 3. 重複的內容： script: 後蓋前
93. extends: 1. 把繼承來源的內容抄⼀次 2. 可以允許既成多個來源 3. 重複的內容： script: 後蓋前 variables: 取合集後蓋前
94. 4.利⽤Anchors 整理可複⽤的 script
95. 4.利⽤Anchors 整理可複⽤的 script
96. 4.利⽤Anchors 整理可複⽤的 script
97. Anchors 1. 可允許多⾏ 2. 可在 script 中使⽤
98. Anchors 1. 可允許多⾏ 2. 可在 script 中使⽤
99. Anchors 1. 可允許多⾏ 2. 可在 script 中使⽤
100. 5. 建立同類⼯ 作模板 tmpl
101. 5. 建立同類⼯ 作模板 tmpl
102. 5. 建立同類⼯ 作模板 tmpl extends: 可以允許多層次繼承
103. 6. 擷取為 變數控制
104. 6. 擷取為 變數控制
105. 6. 擷取為 變數控制
106. 6. 擷取為 變數控制
107. 7. 整理⾄ Template 中
108. 7. 整理⾄ Template 中
109. 7. 整理⾄ Template 中
110. 8. 使⽤ Parallel Matrix
111. 8. 使⽤ Parallel Matrix
112. 8. 使⽤ Parallel Matrix
113. 8. 使⽤ Parallel Matrix
114. 9. 擷取 Parallel Matrix 為模板供 繼承
115. 9. 擷取 Parallel Matrix 為模板供 繼承
116. 9. 擷取 Parallel Matrix 為模板供 繼承 entends: 允許既成多個來源
117. 10. 拆分檔案利 ⽤ include 載入
118. 10. 拆分檔案利 ⽤ include 載入
119. 10. 拆分檔案利 ⽤ include 載入 ├── .gitlab-ci.yml └── templates ├── common.yml ├── de f i ne.yml ├── lint.yml ├── test.yml └── deploy.yml
120. 10. 拆分檔案利 ⽤ include 載入 ├── .gitlab-ci.yml └── templates ├── common.yml ├── de f i ne.yml ├── lint.yml ├── test.yml └── deploy.yml
121. 10. 拆分檔案利 ⽤ include 載入 ├── .gitlab-ci.yml └── templates ├── common.yml ├── de f i ne.yml ├── lint.yml ├── test.yml └── deploy.yml
122. 10. 拆分檔案利 ⽤ include 載入 ├── .gitlab-ci.yml └── templates ├── common.yml ├── de f i ne.yml ├── lint.yml ├── test.yml └── deploy.yml
123. 10. 拆分檔案利 ⽤ include 載入 ├── .gitlab-ci.yml └── templates ├── common.yml ├── de f i ne.yml ├── lint.yml ├── test.yml └── deploy.yml
124. 10. 拆分檔案利 ⽤ include 載入 ├── .gitlab-ci.yml └── templates ├── common.yml ├── de f i ne.yml ├── lint.yml ├── test.yml └── deploy.yml
125. 10. 拆分檔案利 ⽤ include 載入 ├── .gitlab-ci.yml └── templates ├── common.yml ├── de f i ne.yml ├── lint.yml ├── test.yml └── deploy.yml
126. 10. 拆分檔案利 ⽤ include 載入 ├── .gitlab-ci.yml └── templates ├── common.yml ├── de f i ne.yml ├── lint.yml ├── test.yml └── deploy.yml
127. include: - local: 同專案下的檔案 - remote: URL 連結 - template: GitLab 官⽅提供的模板 - f i le: 來⾃其他專案的檔案
128. 11. 建立共⽤流 程供其他專案 使⽤
129. include: f i le introduced in GitLab 11.7 include multi f i le from a project in GitLab 13.6 11. 建立共⽤流 程供其他專案 使⽤
130. 變更及覆寫參數 11. 建立共⽤流 程供其他專案 使⽤
131. 變更及覆寫參數 11. 建立共⽤流 程供其他專案 使⽤
132. 12. 從模板選 擇需要的⼯作
133. 12. 從模板選 擇需要的⼯作
134. Recap • 利⽤ default stage 整理所有 job 都共同的 • 利⽤ .hidden 建立模板⼯作，並⽤ extends 繼承使⽤ • 利⽤ Anchors 整理可重複使⽤的 Script • 萃取出變數 • 利⽤ Parallel Matrix 功能 • 建立多專案共⽤流程 • Sample Source Code: https://bit.ly/3mTiBgJ
135. Sample Code • https://bit.ly/3mTiBgJ
136. 歡迎加入 GitLab Hero
137. Q&A