Message-ID: <14582e1f-4a34-46b8-a20a-ad653930ecb4@cloud.gatewaynet.com>
Date: Mon, 10 Mar 2025 15:35:54 +0200
MIME-Version: 1.0
Subject: Re: log repl exercise monitor setup
To: pgsql-admin@lists.postgresql.org
References: <7c760edb024dc59e92c7c624ae5e00052c8becde.camel@ucll.be>
Content-Language: en-US
From: Achilleas Mantzios - cloud <a.mantzios@cloud.gatewaynet.com>
In-Reply-To: <7c760edb024dc59e92c7c624ae5e00052c8becde.camel@ucll.be>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Archived-At: 
 <https://www.postgresql.org/message-id/14582e1f-4a34-46b8-a20a-ad653930ecb4%40cloud.gatewaynet.com>
Precedence: bulk


On 3/10/25 14:49, Wim Bertels wrote:
> Hello,
>
> i would like to setup a mass test environment for 250+ students
>
> for example a setup with two clusters that we manage: cluster PUB and
> cluster SUB
>
> goal: on cluster PUB each student has its own database
> mirrored/replicated to his own database at cluster SUB
>
> starting point:
> * the databases on cluster PUB already exist with content
> * the empty databases on cluster SUB already exist (without any
> content)
> * no publication (and thus subscription is already running)
>
> so the exercise for each student would be:
> 1. create publication
> 2. mirror the structure (backup/restore)
> 3. create subscription
> 4. test replication (with dml: insert/..)
Sounds like fun project! My daughter studying in 
https://www.ece.tuc.gr/en/home also had access to a automated personal 
PostgreSQL DB while she was attending the DB course. They only learned 
SQL tho, no admin stuff, it is cool that you guys are teaching this.
> So what would be your suggestion to setup an easy monitoring setup to
> check that the student succeeded in this exercise
I'd create custom scripts with count(*) from all tables, sorted by table 
name and verifying those two match and that the #lines match the 
expected number of tables.
>
> background: vanilla debian with postgres apt repository, preferable
> with software from these repositories for security and maintenance
> reasons
>
> (at the moment i check with some handmade scripts that query the
> cluster and the logs)
>
> mvg,
> Wim