On 5/10/19 3:35 PM, Tom Lane wrote:
> Andres Freund <andres@anarazel.de> writes:
>> On 2019-05-10 11:38:57 -0400, Tom Lane wrote:
>>> I am wondering if, somehow, the stack depth limit seen by the postmaster
>>> sometimes doesn't apply to its children. That would be pretty wacko
>>> kernel behavior, especially if it's only intermittently true.
>>> But we're running out of other explanations.
>> I wonder if this is a SIGSEGV that actually signals an OOM
>> situation. Linux, if it can't actually extend the stack on-demand due to
>> OOM, sends a SIGSEGV. The signal has that information, but
>> unfortunately the buildfarm code doesn't print it. p $_siginfo would
>> show us some of that...
>> Mark, how tight is the memory on that machine? Does dmesg have any other
>> information (often segfaults are logged by the kernel with the code
>> IIRC).
> It does sort of smell like a resource exhaustion problem, especially
> if all these buildfarm animals are VMs running on the same underlying
> platform. But why would that manifest as "you can't have a measly two
> megabytes of stack" and not as any other sort of OOM symptom?
>
> Mark, if you don't mind modding your local copies of the buildfarm
> script, I think what Andres is asking for is a pretty trivial addition
> in PGBuild/Utils.pm's sub get_stack_trace:
>
> my $cmdfile = "./gdbcmd";
> my $handle;
> open($handle, '>', $cmdfile) || die "opening $cmdfile: $!";
> print $handle "bt\n";
> + print $handle "p $_siginfo\n";
> close($handle);
>
>
I think we'll need to write that as:
print $handle 'p $_siginfo',"\n";
As you have it written perl will try to interpolate a variable called
$_siginfo.
cheers
andrew