| From: | Harrison Booth <harrisontbooth(at)gmail(dot)com> |
|---|---|
| To: | Bryan Green <dbryan(dot)green(at)gmail(dot)com> |
| Cc: | Andres Freund <andres(at)anarazel(dot)de>, Thomas Munro <thomas(dot)munro(at)gmail(dot)com>, pgsql-hackers(at)lists(dot)postgresql(dot)org |
| Subject: | Re: [PATCH] Fix orphaned backend processes on Windows using Job Objects |
| Date: | 2026-10-16 19:01:21 |
| Message-ID: | CAAH-eSVh63+_OwacE8dkhc+VZkAwLcScumA-ChrReeodFmDuFw@mail.gmail.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Regarding reproduction, the two attachments contain a local rebase of
your public v2, followed by my anonymous-job supplement. The
prerequisite retains the Job logic and recovery test - it is not a new
author revision.
The exact 158b6d5109fa source passed ten server crash/restart cycles
per architecture after clean native Windows x64 and ARM64 builds.
The prerequisite still attempts nested assignment while its comment
says nesting is skipped. Could the comment describe that attempt, and
the access-denied report preserve the actual failure code without
assuming the outer
supervisor handles cleanup? Anonymous naming doesn't establish that
policy or prevent privileged handle duplication.
Best,
Harrison
| Attachment | Content-Type | Size |
|---|---|---|
| 0001-Use-Windows-Job-Objects-to-prevent-orphaned-child-pr.patch | text/x-patch | 14.6 KB |
| 0002-Use-an-unnamed-job-for-PostgreSQL-orphan-cleanup.patch | text/x-patch | 2.2 KB |
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Manu | 2026-10-16 19:55:19 | Re: [PG19] Promotion hangs behind a backend blocked on a stuck client |
| Previous Message | Harrison Booth | 2026-10-16 18:28:35 | Re: [PATCH] pg_upgrade: handle failed nonblocking connects on Windows |