Re: SIGSEGV from START_REPLICATION 0/XXXXXXX in XLogSendPhysical () at walsender.c:2762

Hi,

On 2020-06-24 15:41:14 -0400, Alvaro Herrera wrote:
> On 2020-Jun-24, Robert Haas wrote:
> 
> > So really I think this turns on #1: is it plausible
> > that people are using this feature, however inadvertent it may be, and
> > is it potentially useful? I don't see that anybody's made an argument
> > against either of those things. Unless someone can do so, I think we
> > shouldn't disable this.
> 
> People (specifically the jdbc driver) *are* using this feature in this
> way, but they didn't realize they were doing it.  It was an accident and
> they didn't notice.

As I said before, I've utilized being able to do both over a single
connection (among others to initialize a logical replica using a base
backup). And I've seen at least one other codebase (developed without my
input) doing so. I really don't understand how you just dismiss this
without any sort of actual argument.  Yes, those uses can be fixed to
reconnect with a different replication parameter, but that's code that
needs to be adjusted and it requires adjustments to pg_hba.conf etc.

And obviously you'd lock out older versions of jdbc, and possibly other
drivers.

Obviously we should allow more granular permissions here, I don't think
anybody is arguing against that.

Greetings,

Andres Freund