jumpserver/jumpserver · critical · DatabaseError

Cannot connect to database

Error message

Cannot connect to database

What it means

Raised by OracleClient.commit when self._conn is None — i.e. commit() was called on a client whose JDBC/DB connection was never established or has been lost. The class guards execute paths with fail_json, but commit() raises DatabaseError directly because there is no connection object to commit on.

Source

Thrown at apps/libs/ansible/modules_utils/oracle_common.py:99

        result, error = None, None
        try:
            self.cursor.execute(sql, params)
            sql_header = self.cursor.description or []
            column_names = [description[0].lower() for description in sql_header]
            if column_names:
                result = [dict(zip(column_names, row)) for row in self.cursor]
                result = result[0] if len(result) == 1 else result
            else:
                result = None
        except DatabaseError as err:
            error = err
        if exception_to_fail and error:
            self.module.fail_json(msg='Cannot execute sql: %s' % to_native(error))
        return result, error

    def commit(self):
        if self._conn is None:
            raise DatabaseError('Cannot connect to database')
        self._conn.commit()

    def close(self):
        try:
            if self._cursor:
                self._cursor.close()
            if self._conn:
                self._conn.close()
        except:
            pass

View on GitHub (pinned to 6ec464fabd)

Solutions

1. Check the connect step's result/error before calling commit; abort on connection failure
2. Ensure connect() succeeded (self._conn is not None) — inspect why the connection is missing (credentials, listener, network)
3. Structure the flow as connect -> execute -> commit -> close with early exit on the first failure

Example fix

# before
client.connect()
result, err = client.execute(sql)
client.commit()  # DatabaseError if connect failed silently

# after
client.connect()
if client._conn is None:
    module.fail_json(msg='Oracle connection failed')
result, err = client.execute(sql)
if not err:
    client.commit()

if getattr(client, '_conn', None) is None:
    module.fail_json(msg='Oracle connection not established; aborting before commit')

null

try:
    client.commit()
except DatabaseError as e:
    module.fail_json(msg=f'Commit failed: {e}')

When it happens

Trigger: Calling commit() after connect() failed but the failure was swallowed (exception_to_fail=False or error ignored); calling commit() after close() or after the underlying connection dropped; the main flow reaching the commit step for user_add/user_change_password without a prior successful connect.

Common situations: Oracle asset with wrong host/port/credentials where connect errors were logged but execution continued; network drop mid-task before commit; code paths that call commit() unconditionally after a try/except that swallowed the connect exception.

Related errors

• InvalidElasticsearch
• Password is required for 'password' authentication type
• Unsupported authentication type: {auth_type}
• SSH shell channel is not available
• SSH shell channel closed during handshake

AI-assisted analysis of jumpserver/jumpserver@6ec464fabd (2026-08-28). Data as JSON: /api/errors/3be64546d9379e32. Report an issue: GitHub.