Jenkins build disappears without trace

Summary

Automatic builds in Jenkins are performed, are visible while they are running, but afterwards there is no trace whatsoever of that build ID.

Steps to reproduce

We are running https://ci.eclipse.org/4diac/job/FORTE_DEVELOP_via_FBE/ automatically once per day. While it is running, the build can be accessed and there is live console output exactly as it should be. For example, today at around 8:00 CET build number 681 was running. I could see its output, which was exactly like I expected it to be.

But now, there is no trace of build 681. No failed build, no successful build, nothing. The same happened yesterday, so I suspect it can be reproduced by looking at the build each morning.

What is the current bug behavior?

The build disappears as if it never happened.

What is the expected correct behavior?

The build should be noted as either failed, unstable, or successful build (actually, successful right now).

Relevant logs and/or screenshots

From the Jenkins overview page:

Last build (#676), 4 days 17 hr ago
Last stable build (#676), 4 days 17 hr ago
Last successful build (#676), 4 days 17 hr ago
Last failed build (#670), 6 days 17 hr ago
Last unstable build (#666), 10 days ago
Last unsuccessful build (#672), 5 days 17 hr ago
Last completed build (#676), 4 days 17 hr ago

But this morning, build 681 has been running. Yesterday's build is also missing.

Priority

• Urgent
• High
• Medium
• Low

Severity

• Blocker
• Major
• Normal
• Low

Impact

This prevents regular build testing of multiple Eclipse 4diac subprojects.

changed the description

added IT-prioritymedium IT-severitymajor service:jenkins statetodo team:releng labels

assigned to @fgurr

We are looking into it.

We can reproduce the issue. Build #683 vanished in the Jenkins WebUI after the build was complete. It is stored on disk and everything is there except the build.xml file which contains a lot of metadata.

I have not seen this issue before, so I will need to do some more testing.

Different job on the same CI instance works fine.

marked this issue as related to #5733 (closed)