Add upgrade test for spiderpool #3323
Conversation
ty-dc
force-pushed
the
ci/upgrade_test
branch
11 times, most recently
from
9d54e5d
to
6ddefba
Compare
|
wait #3322 |
ty-dc
force-pushed
the
ci/upgrade_test
branch
14 times, most recently
from
e94cb76
to
c3f2fd7
Compare
ty-dc
force-pushed
the
ci/upgrade_test
branch
4 times, most recently
from
7b23b35
to
9312d4d
Compare
ty-dc
added
release/none
ty-dc
force-pushed
the
ci/upgrade_test
branch
2 times, most recently
from
39215c6
to
87d2a8b
Compare
|
|
||
| workflow_dispatch: | ||
| inputs: | ||
| ref: |
There was a problem hiding this comment.
如果是手动触发的，那么应该是 手动指定某个老版本、或者默认上一个版本
| KUBECONFIG_PATH: /home/runner/work/spiderpool/spiderpool/test/.cluster | ||
| on: | ||
| schedule: | ||
| - cron: "0 20 * * *" |
There was a problem hiding this comment.
如果是自动触发的，我认为 可以是 当前 main 分支的版本的 上一个 最新版本，即可
| fail-fast: false | ||
| matrix: | ||
| # Synchronise with the latest releases of each version | ||
| version: [release-v0.7, release-v0.8] |
There was a problem hiding this comment.
我认为不需要 手动维护 这个，也不需要覆盖很多老的，只要测试到 前后两个版本能顺利升级
There was a problem hiding this comment.
维护的是 release-v0.7、release-0.8、release-0.9 是否应该考虑这几个版本都能升级到最新版本？
There was a problem hiding this comment.
我不认为 这里需要人工维护是个号注意，如果CI 本身证明了 7-8 顺利，8-9 顺利，可能就意味着 7-9顺利 或者有 7-8-9的升级通道
即使是社区的大型项目也不保证 跨多个版本能升级
There was a problem hiding this comment.
并且这里 没有覆盖 v0.7.0 - v0.7.1 ， v0.7.5 - v0.8.0 这些升级情况
所以我认为 考虑 当前代码分支的 上一个最新发布，是最基本的诉求，当前这个诉求还没满足
There was a problem hiding this comment.
并且这里 没有覆盖 v0.7.0 - v0.7.1 ， v0.7.5 - v0.8.0 这些升级情况 所以我认为 考虑 当前代码分支的 上一个最新发布，是最基本的诉求，当前这个诉求还没满足
1
| ref: ${{ needs.get_ref.outputs.ref }} | ||
| fetch-depth: 0 | ||
|
|
||
| - name: Move kubeconfig to branch ${{ needs.get_ref.outputs.ref }} |
There was a problem hiding this comment.
这里可以替换 新老版本的 变量名，便于 代码阅读
oldVersion
upgradeVersion
| .PHONY: helm_upgrade_spiderpool | ||
| helm_upgrade_spiderpool: | ||
| @echo -e "\033[35m [helm upgrade spiderpool] \033[0m" | ||
| kubectl delete po -n $(RELEASE_NAMESPACE) spiderpool-init --kubeconfig $(E2E_KUBECONFIG) || true ;\ |
There was a problem hiding this comment.
这个是升级的必须动作么 ？ 有体现在 文档中 ？
There was a problem hiding this comment.
文档里说 “删除 spiderpool-init Pod（可选）-- 如果您的环境有 spiderpool-init Pod (默认其不是一定有么) ---- 将其删除，否则在 helm upgrade 时，会失败 (没解释为什么 有 会导致失败) ”
那用户如何判断 是否 可选
| @echo -e "\033[35m [helm upgrade spiderpool] \033[0m" | ||
| kubectl delete po -n $(RELEASE_NAMESPACE) spiderpool-init --kubeconfig $(E2E_KUBECONFIG) || true ;\ | ||
| HELM_OPTION="";\ | ||
| HELM_OPTION+=" --set spiderpoolController.replicas=1 " ; \ |
There was a problem hiding this comment.
既然是升级，为什么不是标准的 --resuse-values ，还需要设置这么多
There was a problem hiding this comment.
之前考虑的跨版本升级
有些新增 crd ，所以要删下 spiderpool-init
其次，如果默认是 --resuse-values ，很多情况 Pod 起不来。
如：make setup 设置的 controller pod 是两个副本，如果 --resuse-values ，升级后 controller Pod 还是两个副本，那么会出现 端口占用，新的 controller pod 无法运行。
There was a problem hiding this comment.
我认为这部分代码就是和标准的升级文档是一致的
（1）有些新增 crd ，所以要删下 spiderpool-init ---- 如果是标准流程，升级文档有体现么
（2）--resuse-values ，很多情况 Pod 起不来 --- 那客户怎么办 ？ 升级文档如何体现
ty-dc
force-pushed
the
ci/upgrade_test
branch
2 times, most recently
from
d347280
to
4b3bf28
Compare
ty-dc
force-pushed
the
ci/upgrade_test
branch
2 times, most recently
from
bcfd18b
to
aa60c6a
Compare
ty-dc
force-pushed
the
ci/upgrade_test
branch
2 times, most recently
from
ca16cb0
to
2b2b103
Compare
Signed-off-by: tao.yang <tao.yang@daocloud.io>
Thanks for contributing!
What type of PR is this?
What this PR does / why we need it:
Which issue(s) this PR fixes:
Fixes #
Special notes for your reviewer: