On 26.11.2014 23:26, Peter Geoghegan wrote:
> On Wed, Nov 26, 2014 at 2:00 PM, Andrew Dunstan <andrew@dunslane.net> wrote:
>> The client's question is whether this is not a bug. It certainly seems like
>> it should be possible to plan a query without chewing up this much memory,
>> or at least to be able to limit the amount of memory that can be grabbed
>> during planning. Going from humming along happily to OOM conditions all
>> through running "explain <somequery>" is not very friendly.
>
>
> Have you tried this with a "#define SHOW_MEMORY_STATS" build, or
> otherwise rigged Postgres to call MemoryContextStats() at interesting
> times?
FWIW, this does the trick on a regular build:
gdb -batch -x gdb.cmd -p $PID
where gdb.cmd is a file with a single line:
p MemoryContextStats(TopMemoryContext)
Just execute it at the interesting moment when a lot of memory is consumed.
Tomas
