On 4/10/17 14:10, Josh Berkus wrote: > On 04/10/2017 09:04 AM, Justin Clift wrote: >> On 10 Apr 2017, at 16:42, Peter Eisentraut <peter.eisentraut@2ndquadrant.com> wrote: >>> On 4/8/17 04:50, Michael Banck wrote: >>>> "Client-Side Connection-Failover" >>> >>> I don't think the libpq feature can be considered "failover". It just >>> picks from a list of end points on the initial connection attempt. It >>> does not discover connection failures and reconnect. >> >> Yeah, after reading Roberts blog post, I'd also agree that "failover" >> is a bad/misleading (in the "automatic" sense) description for it. > > Alternate phrasing? > > I do want to get this in the "extended list" of features, just because > we want driver authors to add support for it.
I'm not sure what you mean by this. If the driver is based on libpq, they get it for free. If the driver is not based on libpq, they have to implement it themselves, but don't need to wait for libpq before doing so.
And libpq already copied the idea from jdbc, did it not? So at least one non-libpq driver already has it.