Re: On-demand running query plans using auto_explain and signals

On Fri, Sep 18, 2015 at 6:59 AM, Pavel Stehule <pavel(dot)stehule(at)gmail(dot)com> wrote:
> I am afraid so it has not simple and nice solution - when data sender will
> wait for to moment when data are received, then we have same complexity like
> we use shm_mq.
>
> Isn't better to introduce new background worker with responsibility to clean
> orphaned DSM?

That won't work, or at least not easily. On Windows, the DSM is
cleaned up by the operating system as soon as nobody has it mapped.

Frankly, I think you guys are making this out to be way more
complicated than it really is. Basically, I think the process being
queried should publish a DSM via a slot it owns. The recipient is
responsible for reading it and then notifying the sender. If a second
process requests data before the first process reads its data, the
second process can either (1) become an additional reader of the
already-published data or (2) wait for the first process to finish,
and then send its own inquiry.

There are some problems to solve here, but they hardly seem impossible.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company