featzhang commented on PR #28096:
URL: https://github.com/apache/flink/pull/28096#issuecomment-4368866909

   ## 🔁 Azure CI re-triggered — build #74758 was a well-known `release-1.20` 
infrastructure flaky
   
   I just pushed an empty commit to kick a fresh Azure CI run. Sharing the 
root-cause evidence here so reviewers don't have to dig through the Azure logs 
themselves.
   
   ### ❌ Build #74758 failure location
   
   - **Stage** : `CI build (custom builders)` → **Job** `test_ci connect` → 
**Task** `Test - connect`
   - **Failing test** : `MinioTestContainerTest` in 
`flink-filesystems/flink-s3-fs-base`
   - **Result** : Container startup failed → no `TEST-*.xml` emitted → 
`PublishTestResults` also marked failed
   
   ### 🔍 Exact stack trace (from Azure log `_build/results?buildId=74758`)
   
   ```
   org.testcontainers.containers.ContainerLaunchException:
       Container startup failed for image 
minio/minio:RELEASE.2022-02-07T08-17-33Z
   
   Caused by: com.fasterxml.jackson.databind.exc.InvalidFormatException:
       Can not deserialize value of type java.util.Date from String
       "2022-10-27T14:57:17.758157913+08:00": not a valid representation
       (Failed to parse Date value '2022-10-27T14:57:17.758157913+0800':
        while it seems to fit format 'yyyy-MM-dd'T'HH:mm:ss.SSSZ', parsing 
fails)
       at org.testcontainers.shaded.com.fasterxml.jackson.databind.exc
           .InvalidFormatException.from(InvalidFormatException.java:74)
   ```
   
   ### 🧭 Why this is unrelated to this PR
   
   | Aspect | Value |
   | --- | --- |
   | Files touched by this PR | 4 files, all under 
`flink-runtime-web/web-dashboard/` (TS / HTML / LESS) |
   | Failing module | `flink-filesystems/flink-s3-fs-base` |
   | Relationship | None — the backport does not touch Java, Maven, Docker, or 
testcontainers config |
   
   ### 📐 Why the flaky is specific to `release-1.20`
   
   The testcontainers version pinned on `release-1.20` bundles an older Jackson 
that only accepts the 3-digit-millisecond pattern `yyyy-MM-dd'T'HH:mm:ss.SSSZ`. 
The Azure agents' Docker daemon intermittently returns the state JSON with 
9-digit nanosecond precision (e.g. `...758157913+08:00`), which trips the 
deserializer and aborts container startup. `master` has already moved to a 
newer testcontainers that tolerates nanoseconds; this fix has not been 
backported to 1.20.
   
   ### ✅ Request
   
   If build #74xxx on commit `2d64000c` passes everything except this same 
`MinioTestContainerTest`, I'd kindly ask reviewers to treat it as a known flaky 
and proceed with the merge. Happy to rebase / re-push if any committer prefers 
to wait for a green run.
   
   /cc committers who have touched this area recently — apologies for the noise 
and thanks for the review!
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to