On Thu, Aug 31, 2017 at 2:13 PM, Tom Lane wrote:
> Yeah, it is different. What I'm looking at is that nodeGather does
> DestroyTupleQueueReader as soon as it's seen EOF on a given tuple queue.
> That can't save any worker cycles. The reason seems to be that it wants
> to collapse its array of Tu
Robert Haas writes:
> On Thu, Aug 31, 2017 at 11:09 AM, Tom Lane wrote:
>> (With this patch,
>> there are no callers of shm_mq_get_queue(); should we remove that?)
> May as well. I can't remember any more why I did shm_mq_detach() that
> way; I think there was someplace where I thought that the
On Thu, Aug 31, 2017 at 11:09 AM, Tom Lane wrote:
> I complained a couple weeks ago that nodeGatherMerge looked like it
> leaked a lot of memory when commanded to rescan. Attached are three
> proposed patches that, in combination, demonstrably result in zero
> leakage across repeated rescans.
Go
I complained a couple weeks ago that nodeGatherMerge looked like it
leaked a lot of memory when commanded to rescan. Attached are three
proposed patches that, in combination, demonstrably result in zero
leakage across repeated rescans.
The first thing I noticed when I started digging into this wa
