After the recent RHEL OS patching on dolnxdb099pv, the application is no longer able to connect to the PostgreSQL database on port 6526. The application fails with the following error:
psycopg2.OperationalError: connection to server at "dolnxdb099pv" (200.24.1.108), port 6526 failed: server closed the connection unexpectedly This probably means the server terminated abnormally before or while processing the request.
Important detail: when I log in to the RHEL server directly, I am able to connect to the database and query it without any issue. The database instance itself is up and healthy. The failure is only on connections coming in from the application host.
Since the connection is accepted and then dropped (rather than refused), this points to something in the network or TLS path rather than the database engine — most likely a change introduced by the patching, such as:
- firewalld / iptables rules reinstated or reset after the patch - an OS-level security agent or inspection layer now intercepting the connection - an OpenSSL / TLS library update changing the negotiation behaviour for remote clients - SELinux policy reapplied after the patch
Could you please check from your side whether the patching activity changed the host firewall, SELinux state, or any security/network agent configuration on dolnxdb099pv? I will continue checking the PostgreSQL server log and pg_hba configuration in parallel and share anything relevant.
This is currently blocking the application, so any quick input would be appreciated.
