Andres Freund <andres@anarazel.de> writes:
> On 2022-11-29 20:44:34 -0500, Tom Lane wrote:
>> It's also strange that we're apparently running with signals enabled
>> whereever it is that rtld_bind is getting called from. Could it be that
>> sigaction is failing to install the requested signal mask, so that one
>> postmaster signal handler is interrupting another?
> This made me look at pqsignal_pm() / pqsignal() and realize that we wouldn't
> even notice if it failed, because they just return SIG_ERR and callers don't
> check. I don't think that's a likely to be related, but theoretically it could
> lead to some odd situations.
Yeah, I noticed that just now too. But if sigaction() failed,
the signal handler wouldn't get installed at all, which'd lead
to different and more-obvious symptoms. So I doubt that that's
what happened.
regards, tom lane