MIME-Version: 1.0
References: <CA+5UsmmZCFKtp8kuW8AE8WKnkdE=WYvnh3BEC5USHkk96wBxZA@mail.gmail.com>
In-Reply-To: <CA+5UsmmZCFKtp8kuW8AE8WKnkdE=WYvnh3BEC5USHkk96wBxZA@mail.gmail.com>
From: Amritanshu Joshi <amritanshujoshi@gmail.com>
Date: Tue, 1 Jul 2025 14:48:42 -0500
Message-ID: <CA+5Usmkn60jHK2BFHh=oaOAivH7zoabtpD78ia+U1mXPQ1x81A@mail.gmail.com>
Subject: Fwd: Anomalous behavior between two Aurora Postgres RDS clusters
To: pgsql-general@lists.postgresql.org
Content-Type: multipart/alternative; boundary="000000000000afac320638e370c1"
Archived-At: <https://www.postgresql.org/message-id/CA%2B5Usmkn60jHK2BFHh%3DoaOAivH7zoabtpD78ia%2BU1mXPQ1x81A%40mail.gmail.com>
Precedence: bulk

--000000000000afac320638e370c1
Content-Type: text/plain; charset="UTF-8"

Hello,

I have a question that involves two Aurora Postgres RDS clusters.

I have two clusters. With long-running transactions, one of them is able to
create a replication slot when a record is inserted into a table while the
other does not and is blocked. It does create the replication slot when the
transaction is committed. I wanted to know if this behavior is actually an
anomaly or there is a misconfiguration with the parameters that is creating
the replication slot when it should be blocked?

Also, for the long running transactions, the transaction isolation
level is *read
committed *for both the clusters.

Thank you,
Amritanshu

--000000000000afac320638e370c1
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div class=3D"gmail_quote gmail_quote_container"><div dir=
=3D"ltr"><div>Hello,</div><div><br></div><div>I have a question that involv=
es two Aurora Postgres RDS clusters.</div><div><br></div><div>I have two cl=
usters. With long-running transactions, one of them is able to create a rep=
lication slot when a record is inserted into a table while the other does n=
ot and is blocked. It does create the replication slot when the transaction=
 is committed. I wanted to know if this behavior is actually an anomaly or =
there is a misconfiguration with the parameters that is creating the replic=
ation slot when it should be blocked?</div><div><br></div><div>Also, for th=
e long running transactions, the transaction isolation level is <i>read com=
mitted </i>for both the clusters.</div><div><br></div><div>Thank you,</div>=
<div>Amritanshu</div><div><br></div></div>
</div></div>

--000000000000afac320638e370c1--