On Mon, Dec 11, 2017 at 10:07 PM, Craig Ringer <craig@2ndquadrant.com> wrote: > On 12 December 2017 at 12:43, Andres Freund <andres@anarazel.de> wrote: >> On 2017-12-12 11:57:41 +0800, Craig Ringer wrote: >> But that'd have >> the disadvanatage that it possibly would take a while till the >> MemoryContextStats() is executed. Not sure if that's still good enough >> for you? > > Definitely. Sure, it won't be perfect, but it'd be a big improvement on what > we have.
If this would be fine enough, why not giving a shot then? Having to use gdb now on production systems is something sometimes hard to justify to customers. There are also the Windows problems... >> Another question is whether printing to stderr, bypassing proper >> logging!, is good enough for something like this. > > I think the reason it prints to stderr now is that it's intended to run in > OOM situations.
Yep. I am on board with Tom here that this property should not be thrown away.
OK, so I think I'll do pretty much what I outlined above, using stringinfo for the !OOM case and fprintf for the OOM case, via callback, roughly as outlined upthread. I won't bother with elog, it's easy enough if someone cares.
