Re: windows CI failing PMSignalState->PMChildFlags[slot] == PM_CHILD_ASSIGNED - Mailing list pgsql-hackers

From Thomas Munro
Subject Re: windows CI failing PMSignalState->PMChildFlags[slot] == PM_CHILD_ASSIGNED
Date
Msg-id CA+hUKGJgr4msGWt-0iemUVYz7oLs2xgekDGWuVErq52H4=jbmw@mail.gmail.com
Whole thread Raw
In response to Re: windows CI failing PMSignalState->PMChildFlags[slot] == PM_CHILD_ASSIGNED  (Thomas Munro <thomas.munro@gmail.com>)
Responses Re: windows CI failing PMSignalState->PMChildFlags[slot] == PM_CHILD_ASSIGNED
List pgsql-hackers
On Wed, Feb 8, 2023 at 2:52 PM Thomas Munro <thomas.munro@gmail.com> wrote:
> On Wed, Feb 8, 2023 at 2:28 PM Andres Freund <andres@anarazel.de> wrote:
> > TRAP: failed Assert("PMSignalState->PMChildFlags[slot] == PM_CHILD_ASSIGNED"), File:
"../src/backend/storage/ipc/pmsignal.c",Line: 329, PID: 5948
 

I was wondering if commit 18a4a620 might be relevant, as it touched
the management of those slots a few months ago, but then I found a
couple of matches from 2021 in my database of build farm assertion
failures:

https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=fairywren&dt=2021-08-12%2010:38:56
https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=fairywren&dt=2021-09-28%2016:40:49

Fairywren is a Windows + msys2 system, so it doesn't share Cygwin's
signal system, it's running the pure Windows code (though it's GCC
instead of MSVC and has a different libc, it's using our Windows
native code paths and defines WIN32).



pgsql-hackers by date:

Previous
From: "Sho Kato (Fujitsu)"
Date:
Subject: Difference of ecpg japanese translation in PG15
Next
From: "wangw.fnst@fujitsu.com"
Date:
Subject: RE: Perform streaming logical transactions by background workers and parallel apply