Re: [PoC] pg_upgrade: allow to upgrade publisher node

On Wed, Aug 16, 2023 at 4:51 PM Zhijie Hou (Fujitsu)
<houzj.fnst@fujitsu.com> wrote:
>
> 4.
> +int    num_slots_on_old_cluster;
>
> Instead of a new global variable, would it be better to record this in the cluster info ?
>

I was thinking whether we can go a step ahead and remove this variable
altogether. In old cluster handling, we can get and check together at
the same place and for the new cluster, if we have a function that
returns slot_count by traversing old clusterinfo that should be
sufficient. If you have other better ideas to eliminate this variable
that is also fine. I think this will make the patch bit clean w.r.t
this new variable.

--
With Regards,
Amit Kapila.