Mysql Replication: error connecting

[Anon]

Hi I'm having strange yet serious problems with the new replication functions of mysql.

I have set up my /etc/my.cnf file correctly as far as I can see, yet when I start the slave I get the following errors in my error log:

010129 15:54:51 Slave thread: error connecting to master:Can't connect to MySQL
server on '0' (11)(107), retry in 0 sec

Strangely, when I do a SHOW SLAVE STATUS it reports that there is no master host, no master user, no master password and no reconnect period. BUT YET, it has read the replicate-do-db part of the my.cnf file. This is very strange. Can anyone help? I have followed instructions to the letter, and have doubled checked everything time and time again!

Thanks
Andrew Cuthbert

[lawnmowerman]

Can you post the /etc/my.cnf files for the master and slave?

-Rich

[Anon]

Sure:

Master_____________________
[mysqld]
log-bin
server-id=1

Slave______________________
[mysqld]
master-host=212.67.197.79

master-user=repl
master-password=##Password##
replicate-do-db=bentlys
master-connect-retry=300
server-id=2

For security the password has been overwritten above.

Hope you can help!
Thanks,
Andrew

[lawnmowerman]

Okay...everything seems to be fine in your my.cnf files from first glance.

Can you access the master from the slave from the command line?

mysql -u repl --password='##Password##' -h 212.67.197.79 bentlys

-Rich

[Anon]

Hi there,

No, initially I couldn't access the database bentlys. I could login to the server without selecting a database however. I therefore added FILE permissions for the user repl on database bentlys, and stopped and started the slave - but the same error is reported. It doesn't seem to be reading the configuration file correctly as SHOW SLAVE STATUS; returns:

+-------------+-------------+-------------+---------------+-------------+-----+---------------+-----------------+---------------------+
| Master_Host | Master_User | Master_Port | Connect_retry | Log_File | Pos | Slave_Running | Replicate_do_db | Replicate_ignore_db |
+-------------+-------------+-------------+---------------+-------------+-----+---------------+-----------------+---------------------+
| 0 | 0 | 0 | 0 | ns-bin.001 | 0 | No | bentlys | |
+-------------+-------------+-------------+---------------+-------------+-----+---------------+-----------------+---------------------

It has picked up on the replicate_do_db but not the master_**** information. Any further ideas? Thanks for your help by the way.

Andrew

[lawnmowerman]

This may be redundant information, but from what I understand, "SHOW MASTER STATUS" must be run when connected to the master, and "SHOW SLAVE STATUS" must be run when connected to the slave.

Also, make sure that the 'repl' has privs to access the database on the master.

You also have to reload or refresh on the master and slave after each change you do (at least I did to get my replication working).

I also gave my replicate user full access to the database on the master server, not just FILE privs.

Oh, one more though, what versions of MySQL are the master and slave running?

-Rich