On Tue, Aug 14, 2012 at 05:56:39PM -0400, Tom Lane wrote:
> Peter Eisentraut <peter_e@gmx.net> writes:
> > On 8/10/12 7:48 PM, Dimitri Fontaine wrote:
> >> What about having single user mode talk fe/be protocol, and talk to it via a UNIX pipe, with pg_upgrade starting
thesingle user backend as a subprocess?
>
> > I think that's essentially equivalent to starting the server on a
> > Unix-domain socket in a private directory. But that has been rejected
> > because it doesn't work on Windows.
>
> > The question in my mind is, is there some other usable way on Windows
> > for two unrelated processes to communicate over file descriptors in a
> > private and secure way?
>
> You're making this unnecessarily hard, because there is no need for the
> two processes to be unrelated.
>
> The implementation I'm visualizing is that a would-be client (think psql
> or pg_dump, though the code would actually be in libpq) forks off a
> process that becomes a standalone backend, and then they communicate
> over a pair of pipes that were created before forking. This is
> implementable on any platform that supports Postgres, because initdb
> already relies on equivalent capabilities.
I think the big question is whether we need to modify every binary that
pg_upgrade executes to underestand this pipe communication method.
-- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB
http://enterprisedb.com
+ It's impossible for everything to be true. +