MIME-Version: 1.0
References: 
 <CAKoxK+6mkk5r3WREDUduu-ZQH0S87x=cddYDMDyvYyfTzDxj1w@mail.gmail.com>
 <CAKoxK+5+LFK4kk0Xay+9Rs8AYj9we0_M=T9md4HP0+rjZH6=uQ@mail.gmail.com>
In-Reply-To: 
 <CAKoxK+5+LFK4kk0Xay+9Rs8AYj9we0_M=T9md4HP0+rjZH6=uQ@mail.gmail.com>
From: Luca Ferrari <fluca1978@gmail.com>
Date: Wed, 29 Oct 2025 09:22:39 +0100
Message-ID: 
 <CAKoxK+5D4szPoPOWvg6aaJLWEKDf5sHQFgNOCJrod+swKJWCSA@mail.gmail.com>
Subject: Re: what is causing different pcp_node_info status
To: pgpool-general@lists.postgresql.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Archived-At: 
 <https://www.postgresql.org/message-id/CAKoxK%2B5D4szPoPOWvg6aaJLWEKDf5sHQFgNOCJrod%2BswKJWCSA%40mail.gmail.com>
Precedence: bulk

On Thu, Oct 23, 2025 at 8:04=E2=80=AFAM Luca Ferrari <fluca1978@gmail.com> =
wrote:
>
> I am still observing the issue, the best situation I come is all nodes
> reported as "up up" when queryed on pg1 (primary) and "waiting up"
> when queried from pg2 and pg3.
> could it be network latency or a similar issue?

As explained clearly here
<https://github.com/pgpool/pgpool2/issues/131#issuecomment-3436197043>
the information about the nodes is collected and kept in shared
memory.
Is there any chance the same process is applied to the pgpool node status?
I observed that even a machine reboot does not seem to make pgpool
report a status better than "waiting up", evne after minutes, and
while replicas are streaming.
Restarting then the pgpool service brings the status "up up", this
makes me think there is something blocking or delaying pgpool to
collect data, and when the service is restarted the node is forced to
"dig" the other node status.
Something I could investigate on?