Thanks for the reply. I agree with your first solution but I feel this is very limiting. I would like be able to connect with a dns name with assurance that it will succeed. I think this is a reasonable feature to request. I have, for example, a DNS Round Robin used for load sharing. It is essencial here that the DNS name authenticates.
For clarity below is the error message I receive when the system goes down. The connections were made perfectely well with dns prior to the mysterious switch - I guess the question remains why is it switching? 040119 12:01:16 Slave I/O thread: error reconnecting to master '[EMAIL PROTECTED]:3336': Error: 'Host '111.111.11.11' is not allowed to connect to this MySQL server' errno: 1130 retry-time: 60 retries: 86400 (Numbers and letter changed intentionally.) -----Original Message----- From: Mikael Fridh To: DePhillips, Michael P; [EMAIL PROTECTED] Sent: 1/23/2004 6:50 PM Subject: Re: Replication Madness I'm not sure I quite follow you here. I think you mean that when the slave connects to the master, sometimes the master does not resolve the address the slave has - thus failing because you don't have grants for the slave's IP address.. Generally I think it's a bad idea to be dependent on a dns lookup, the grant should be for the proper Ip adress(es) instead. A name server look up always have the risk of failing (unless it's set to retry endlessly) You could try a worse solution - put the slave's IP address in the HOSTS file on the master. That way (IF the OS reads the host file before it queries a name server) it will always identify the slave's hostname. But like I said, that's the bad solution I think... Mikael ----- Original Message ----- From: "DePhillips, Michael P" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Friday, January 23, 2004 11:00 PM Subject: Replication Madness > Hi Folks > > > > Every now and again one of my slaves try's to connect to its master via its > ip address as opposed to its dns name. This causes a problem because the ip > address does not have credentials on the master, therefore, connection > refused ->replication ends ->databases out of sync-> angry users-> we all > know the rest. The switch is random both temporally and machine wise. Are > there any ideas on why this is happening or what needs to be done to prevent > it? > > > > Perhaps this is a bug and should be reported as such? > > > > Thanks > > Michael > > -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
