Re: pg_background contrib module proposal

2017-01-27 19:14 GMT+05:00 Peter Eisentraut <peter(dot)eisentraut(at)2ndquadrant(dot)com>:
> I suppose we should decide first whether we want pg_background as a
> separate extension or rather pursue extending dblink as proposed elsewhere.
>
> I don't know if pg_background allows any use case that dblink can't
> handle (yet).
pg_background in it's current version is just a start of a feature.
The question is: are they coherent in desired features? I do not know.
E.g. will it be possible to copy from stdin in dblink and possible
incarnations of pg_background functionality?)

2017-01-27 19:38 GMT+05:00 Robert Haas <robertmhaas(at)gmail(dot)com>:
> For the record, I have no big problem with extending dblink to allow
> this instead of adding pg_background. But I think we should try to
> get one or the other done in time for this release.
+1!
that's why I hesitate between not saying my points and making
controversy...need to settle it somehow.

Parallelism is a "selling" feature, everything has to be parallel for
a decade already (don't we have parallel sequential scan yet?).
It's fine to go with dblink, but dblink docs start with roughly "this
is an outdated substandard feature"(not a direct quote[0)].
What will we add there? "Do not use dblink for linking to databases.
This is the standard for doing concurrency." ?
Please excuse me for exaggeration. BTW, pg_background do not have docs at all.

[0] https://www.postgresql.org/docs/devel/static/dblink.html

Best regards, Andrey Borodin.