On 2018-07-24 14:02:26 -0500, Dimitri Maziuk wrote:
> On 07/24/2018 01:43 PM, Andres Freund wrote:
> > On 2018-07-24 12:22:24 -0500, Dimitri Maziuk wrote:
> >> On 07/24/2018 12:21 PM, Alvaro Herrera wrote:
> >>>
> >>> Are you serious about us trying to diagnose a bug with this description?
> >>
> >> What do you want to know, exactly?
> >
> > A recipe that we can follow and reproduce the issue.
>
> The nutshell version is as I said: I pretty much followed the manual to
> create a test publication/subscription setup and let it run for a few
> weeks. I had to increase a few limits but otherwise everything's at the
> default settings as shipped in the rpms.
Are you really expecting us to be able to reproduce the problem based on
the above description? Our test suites do setup plain replications
setups, and the problem doesn't occur there.
> If you are asking for something other than those two extremes, please
> let me know what it is and I'll be happy to try and provide it.
A command by command reproducer would be good.
> I can send you postgres config files from both nodes and the entire
> database dump -- it's all public domain. I can `zfs send` you the
> snapshot of the entire /var/lib/pgsql/10 as is, too.
Without the genesis, that's probably not as helpful, unfortunately.
Greetings,
Andres Freund