Works well but getting abends after a while

[aniellod]

After working for a few hours, the MVSMF seems to abend. Here's the info should it be useful.

abends.txt

Really nice work.

[mgrossmann]

It looks like this happened due to a race condition. Try to disable mqtt support in the httpd server.

[mgrossmann]

Hi @aniellod

thanks for reporting this and sorry for the late response!

I had a look at the dump you provided. The ABEND SA03 is a secondary ABEND during task termination — meaning the original error triggered a cleanup that also failed. The traceback shows two threads involved: the telemetry thread stuck in an ENQ/DEQ operation on a /zosmf/... resource, and the MQTT processing thread in a selectex wait. This pattern (works fine for hours, then abends) could point to a resource serialization issue, like an ENQ not being properly released on an error path.

Is this still an issue for you? If so, it would help a lot to have:

• HTTPD logs from around the time of the ABEND — especially the last few requests before the crash
• How you're connecting — Zowe Explorer, Zowe CLI, or something else?
• What's happening when it abends — active use or sitting idle?
• Any LOGREC entries if available — the primary ABEND code (before the SA03) would be the key to finding the root cause

Thanks again for trying out mvsMF!

[aniellod]

Thanks for the reply. No worries about timeliness, we do what we can. I've noted the additional information needed and will try to capture it. I was connecting via Zowe Explorer for Visual Studio Code. This makes the entire coding experience so much easier. Regards, Aniello.

…

________________________________ De : Mike Großmann ***@***.***> Envoyé : 11 février 2026 15:05 À : mvslovers/mvsmf ***@***.***> Cc : aniellod ***@***.***>; Mention ***@***.***> Objet : Re: [mvslovers/mvsmf] Works well but getting abends after a while (Issue #5) [https://avatars.githubusercontent.com/u/784509?s=20&v=4]mgrossmann left a comment (mvslovers/mvsmf#5)<#5 (comment)> Hi @aniellod<https://github.com/aniellod> thanks for reporting this and sorry for the late response! I had a look at the dump you provided. The ABEND SA03 is a secondary ABEND during task termination — meaning the original error triggered a cleanup that also failed. The traceback shows two threads involved: the telemetry thread stuck in an ENQ/DEQ operation on a /zosmf/... resource, and the MQTT processing thread in a selectex wait. This pattern (works fine for hours, then abends) could point to a resource serialization issue, like an ENQ not being properly released on an error path. Is this still an issue for you? If so, it would help a lot to have: * HTTPD logs from around the time of the ABEND — especially the last few requests before the crash * How you're connecting — Zowe Explorer, Zowe CLI, or something else? * What's happening when it abends — active use or sitting idle? * Any LOGREC entries if available — the primary ABEND code (before the SA03) would be the key to finding the root cause Thanks again for trying out mvsMF! — Reply to this email directly, view it on GitHub<#5 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AANBZDLSDKWI2HDN3B6AJR34LNAMRAVCNFSM6AAAAACJSMXO5CVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTQOBVGAZDQNBXGY>. You are receiving this because you were mentioned.Message ID: ***@***.***>