Re: Set new system identifier using pg_resetxlog

On 2014-06-18 13:26:37 -0400, Robert Haas wrote:
> My vote is to hold off on this until we've talked about replication
> identifiers and other related topics in more depth.

I really don't understand this. We're *NOT* proposing this patch as an
underhanded way of preempting the discussion of whether/how replication
identifiers are going to be used. We're proposing it because we
currently have a need for the facility and this will reduce the number
of patches we have to keep around after 9.5. And more importantly
because there's several other use cases than our internal one for it.

To settle one more point: I am not planning to propose BDR's generation
of replication identifier names for integration. It works well enough
for BDR but I think we can come up with something better for core. If I
had my current knowledge two years back I'd not have chosen the current
scheme.

Greetings,

Andres Freund

-- Andres Freund                       http://www.2ndQuadrant.com/PostgreSQL Development, 24x7 Support, Training &
Services