Sending SIGABRT to child processes (was Re: Strange failure on mamba)

Andres Freund <andres(at)anarazel(dot)de> writes:
> On 2022-11-17 17:47:50 -0500, Tom Lane wrote:
>> So I'd like to have some way to make the postmaster send SIGABRT instead
>> of SIGKILL in the buildfarm environment. The lowest-tech way would be
>> to drive that off some #define or other. We could scale it up to a GUC
>> perhaps. Adjacent to that, I also wonder whether SIGABRT wouldn't be
>> more useful than SIGSTOP for the existing SendStop half-a-feature ---
>> the idea that people should collect cores manually seems mighty
>> last-century.

> I suspect that having a GUC would be a good idea. I needed something similar
> recently, debugging an occasional hang in the AIO patchset. I first tried
> something like your #define approach and it did cause a problematic flood of
> core files.

Yeah, the main downside of such a thing is the risk of lots of core files
accumulating over repeated crashes. Nonetheless, I think it'll be a
useful debugging aid. Here's a proposed patch. (I took the opportunity
to kill off the long-since-unimplemented Reinit switch, too.)

One thing I'm not too clear on is if we want to send SIGABRT to the child
groups (ie, SIGABRT grandchild processes too). I made signal_child do
so here, but perhaps it's overkill.

regards, tom lane