azure-sdk-for-js: [Key Vault Certificates]The integration test "using beginDeleteCertificate's poller" is not finished until timeout on China cloud
We are running live tests against other clouds like US Gov and Azure China Cloud. The goal is to check whether new azure sdk package work with other clouds or not.
Error Description:
When running the The integration test using beginDeleteCertificate's poller
on China cloud, it will not be finished until timeout. The error happens inconsistently and about one time out of two. The error message is shown as following, for more details please check here
Error Track:
There is something wrong with delete operation. When use poller returned by delete operation to run the method pollUntilDone
, it is not responding until timeout. The process code is shown as following:
https://github.com/Azure/azure-sdk-for-js/blob/main/sdk/keyvault/keyvault-certificates/test/public/CRUD.spec.ts#L423
Expected Behavior: The integration test “using beginDeleteCertificate’s poller” can run successfully stably on China cloud.
@benbp, @ramya-rao-a, @AlexGhiondea, @dw511214992 for notification.
About this issue
- Original URL
- State: closed
- Created 3 years ago
- Comments: 16 (16 by maintainers)
Commits related to this issue
- [KeyVault] - Fix certificate poller test (#19909) **Checklists** - [x] Added impacted package name to the issue description **Issues associated with this PR:** Fixes #18680 **Describe the pr... — committed to Azure/azure-sdk-for-js by maorleger 2 years ago
- [KeyVault] - Fix certificate poller test (#20176) ### Packages impacted by this PR @azure/keyvault-certificates ### Issues associated with this PR Fixes #18680 ### Describe the problem that i... — committed to Azure/azure-sdk-for-js by maorleger 2 years ago
My observation is that sometimes the DELETE call returns 200 OK but doesn’t actually delete the certificate. Then we poll forever waiting for it to show up in the deleted certificates collection. I am engaging the service team with traces. Stay tuned!
So the root cause ended up being the same in both tests - we needed to wait for the certificate create poller to complete its operation.
There are likely other tests that need some attention in sovereign clouds, but I made sure these are passing in CI. For any other tests please file a new issue and I will address it ASAP. This way we don’t make too many changes at once, or make unnecessary changes. Thanks for letting me know about this!
Looks like you figured out the missing line in
tests.yml
@zedy-wj @maorleger whatever branches you are/were testing on for the weekly pipeline need to be updated/rebased against the latest changes. That will fix the
ProvisionerApplicationOid
error.