[VAL-68] [ELIOT BLueprint : Blu-Val Test Case Issue : K8S Conformance] : [sig-apps] Daemon set [Serial] should rollback without unnecessary restarts Created: 18/Oct/19  Updated: 21/Oct/19  Due: 18/Nov/19  Resolved: 21/Oct/19

Status: Done
Project: Validation
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Task Priority: Medium
Reporter: Abhijit Dasgupta Assignee: Daniel Stoica
Resolution: Done Votes: 0
Labels: None
Remaining Estimate: 3 weeks
Time Spent: Not Specified
Original Estimate: 3 weeks

Attachments: File eliot_blueprint_validation_lab_logs_201910180202.tar.gz    

 Description   

While executing K8S conformance test one case if failing in our ELIOT local environment. Please help to check.
Issue Description:
[sig-apps] Daemon set [Serial] should rollback without unnecessary restarts. [Conformance]

 Comments   
Comment by Abhijit Dasgupta [ 21/Oct/19 ]

Hi Cristina,

All K8S test cases of Blu Validation framework is passed. I am closing the Jira Ticket.

Thank you Validation team for helping in this issue.

Regards
Abhijit

Comment by Abhijit Dasgupta [ 21/Oct/19 ]

Hi Cristina ,

Thank you for the update.
Ok, I think since in our cluster we are not untaint the master node (control plane) may be that might be the reason

Will modify and run the test suite once and update with the results.

Thank you for your guidance.

Regards
Abhijit.

Comment by Cristina Pauna [ 18/Oct/19 ]

Hi Abhijit,

 

The test is failing with the error "Conformance test suite needs a cluster with at least 2 nodes". You can open the sonobuoy archive and check plugins/e2e/results/global/e2e.log for the full log.

I see that your cluster has 2 nodes, so it's possible that the issue is with reading the number of the nodes. The code of the test is at [1] and it uses function GetReadySchedulableNodesOrDie() to retrieve the nodes.

Looking at how the function is implemented [2], it looks like both nodes need to be ready and schedulable. Can you manually verify this is the case?

In any case, this doesn't look like it's a validation issue, more likely it's an issue with the cluster.

 

Thanks,

Cristina

[1] https://github.com/kubernetes/kubernetes/blob/v1.15.4/test/e2e/apps/daemon_set.go#L386

[2] https://github.com/kubernetes/kubernetes/blob/v1.15.4/test/e2e/framework/util.go#L2621

 

Generated at Sat Feb 10 06:07:03 UTC 2024 using Jira 9.4.5#940005-sha1:e3094934eac4fd8653cf39da58f39364fb9cc7c1.