On 12/17/2013 3:35 AM, Christophe Carles wrote:
> I contact you about a problem with replication Databases with MySQL.
> We have a main master MySQL database server that contains data unit.
> This server is only accessible internally.
> We have LAMP servers in DMZ (accessible from the Internet) that
> are slaves of the master.
> The configuration is done with the file / etc / my.cnf (the servers are
> under centos)
> Below is a sample configuration of a slave:
> # Disabling symbolic-links is recommended to prevent assorted security
The use of these settings are deprecated. You should only set or change
your replication coordinates using a CHANGE MASTER TO... command.
> We had a web server who had shutdown.
> I make another one (VM) with the necessary services : http, mysql , php,
> etc ...
> When I configured to replicate the master database , I got an error
> message in : / var / log / mysqld.log
> [Note ] Slave I / O thread: Failed reading log event , reconnecting to
> retry , log ' FIRST ' at position 4
> [ERROR ] Error reading packet from server : File '# ./mysqld-bin.000001
> ' not found ( errno: 2) ( server_errno = 29)
> On the master , I also receive an error message in / var / log / mysqld.log
> [ERROR ] Failed to open log ( file ' # ./mysqld-bin.000001 ', errno 2)
> I made several search on google and I've tested several things on slave
> like deletes logs, configuration file my.cnf ...
> I have found a solution applied is clean logs master server with FLUSH
> Because this server is used for replication to other bases, I have not
> tried this command.
> Is one of you have already met this kind of problems?
> Is using FLUSH LOGS you seem relevant in a case like mine?
The problem is not with your master. The problem is how you told your
slave to start replicating.
This new slave is a copy of some data that corresponds to the state that
the master was in at a particular location in its binary logs (a file
name and an offset -or- a GTID value). That is where you needed to have
this slave start replicating from. Because you did not provide the
correct coordinates (or any at all) you attempted to start from the
first position of the first binary log. As you can see, that file no
What you need to know is what binary log coordinates match the state of
the data you created this new VM with. This we cannot tell you as we
have no details of where your data originated or how you created the copy.
MySQL Senior Principal Technical Support Engineer
Oracle USA, Inc. - Hardware and Software, Engineered to Work Together.
Office: Blountville, TN