v3.2.6 Argo UI failed to render pod log

[liuzqt]

Summary

What happened/what you expected to happen?

What version of Argo Workflows are you running?
3.2.6 (we just upgrade it from 3.2.3 today, so I guess this issue might be introduced between 3.2.3 and 3.2.6. We'll try to roll back to 3.2.3 and see if this issue still exist)

Diagnostics

I'm still not able to find a minimal reproducing example, since this issue only happened in some of our pipelines, and not sure which part of the template triggered this issue.

---

Message from the maintainers:

Impacted by this bug? Give it a 👍. We prioritise the issues with the most 👍.

[alexec]

Most likely introduced by this commit:

f5f6899 fix: Add pod name format annotation. Fixes #6962 and #6989 (#6982)

@JPZ13 would you be able to look into this please?

[JPZ13]

Most likely introduced by this commit:

f5f6899 fix: Add pod name format annotation. Fixes #6962 and #6989 (#6982)

@JPZ13 would you be able to look into this please?

On it

[JPZ13]

@liuzqt Can you share what the pod names are in kubernetes for this workflow that are resulting in the error on the UI? I'm working on a fix and need to make sure there's a 1:1 match between the pod names as they appear in k8s and what our UI pod name function returns

Thanks for reporting the issue!

[liuzqt]

@JPZ13 the Pod UI is not working, I tried to find their information from the json meta, hopefully this can help:

"patch-processing-pipeline-ksp78-1623891970": {
                "id": "patch-processing-pipeline-ksp78-1623891970",
                "name": "patch-processing-pipeline-ksp78.retriable-map-authoring-initializer",
                "displayName": "retriable-map-authoring-initializer",
                "type": "Pod",
                "templateScope": "local/",
                "phase": "Succeeded",
                "boundaryID": "patch-processing-pipeline-ksp78",
                "startedAt": "2022-01-20T01:45:35Z",
                "finishedAt": "2022-01-20T01:45:39Z",
                "progress": "1/1",
                "resourcesDuration": {
                    "cpu": 3,
                    "memory": 3
                },
                # ...

"patch-processing-pipeline-ksp78-1769534216": {
                "id": "patch-processing-pipeline-ksp78-1769534216",
                "name": "patch-processing-pipeline-ksp78.NODE-RETRIABLE-MAP-AUTHORING(1).NODE-GENERATE-EMLANE.pre-emlane-map-builder",
                "displayName": "pre-emlane-map-builder",
                "type": "Pod",
                "templateRef": {
                    "name": "map-builder-task",
                    "template": "map-builder"
                },
                "templateScope": "namespaced/node-generate-emlane-task",
                "phase": "Failed",
                "boundaryID": "patch-processing-pipeline-ksp78-242024017",
                "message": "Error (exit code 1)",
                "startedAt": "2022-01-20T01:53:05Z",
                "finishedAt": "2022-01-20T01:53:08Z",
                "progress": "1/1",
                "resourcesDuration": {
                    "cpu": 4,
                    "memory": 4
                },
                # ...

[JPZ13]

That is super helpful @liuzqt. Thank you for the extra info. I'll tag you on the PR for the fix once it's ready

[liuzqt]

@JPZ13 sorry actually I'm just an Argo user in our company and I have to ask our Argo team to help verify this fix later. I'll update once I get feedback from them.

[JPZ13]

@JPZ13 sorry actually I'm just an Argo user in our company and I have to ask our Argo team to help verify this fix later. I'll update once I get feedback from them.

Sounds good. Thank you @liuzqt

[terrytangyuan]

@JPZ13 I am in touch with @liuzqt’s team. Let me know if there’s any other information you need.

[JPZ13]

Thanks @terrytangyuan. I'm looking for the mapping of the json meta of NodeStatuses that @liuzqt provided above to the pod names as they appear in the kubernetes cluster. Also, if they find a minimal reproducible example, that would be super helpful too