Re: incorrect information ?

[kostya kote]

ok. this is my code written as UnitTest

public class testFailover extends TestCase {
private String driver;
private Properties prop;
private String url;
private Connection conn = null;

public void loopOp() throws SQLException {

try {

conn = getConnection();
// conn.setAutoCommit(false);

while (true) {
try {
doBusinessOp();
Thread.sleep(1000);
} catch (Exception e) {
e.printStackTrace();
}
}
} finally {
conn.close();
}
}

public void doBusinessOp() throws SQLException {
PreparedStatement stmt = null;
ResultSet rs = null;

try {
//
String query = "SELECT * FROM mysql.user";
stmt = conn.prepareStatement(query);

rs = stmt.executeQuery(query);
int i = 0;
while (rs.next()) {
Config.logger.info("Host" + (i++) + ":" + rs.getString("Host"));
}

rs.close(); rs = null;
stmt.close(); stmt = null;
conn.commit();
Config.logger.info("transaction completed!");
} catch (Exception sqlEx) {
Config.logger.info("Error:"+sqlEx.getMessage());
}
}

private Connection getConnection() throws SQLException {

Connection c = null;
prop = new Properties();

prop.setProperty("user", "root");
prop.setProperty("password", "");
prop.setProperty("autoReconnect", "true");
// prop.setProperty("queriesBeforeRetryMaster", "10");
prop.setProperty("secondsBeforeRetryMaster", "10");
prop.setProperty("roundRobinLoadBalance", "false");
// prop.setProperty("failOverReadOnly", "false");

url = "jdbc:mysql://10.50.1.7, 10.50.1.8/company_db";
driver = "org.gjt.mm.mysql.Driver";

try {
Class.forName(driver); // init
} catch (ClassNotFoundException e) {
Config.logger.error("No class found for driver = '" + driver + "'. Exception: " + e);
throw new RuntimeException(e); // raise exeption
}
c = DriverManager.getConnection(url, prop);


return c;
}
}

as I said, I have 2 sql nodes resided at 10.50.1.7 and 10.50.1.8 correspondingly.
when I broke connection with a frist sql node I havent recoonect to second as mentioned in documentation and other sources. If client application coonects only to one server and this server is down that we have single point of failure. isnt it ?



Edited 1 time(s). Last edit at 09/19/2005 08:44AM by kostya kote.