-->-----Original Message----- -->From: Mark T. Dame [mailto:[EMAIL PROTECTED] -->Sent: Friday, August 15, 2003 1:38 PM -->To: MySQL Mailing List -->Subject: Re: Bi-directional Replication --> -->On 08/15/03 16:01, Andreas wrote: -->> -->> I did this some time ago. -->> -->> Host A and B -->> You just configure host A as master and host B as slave of host A. -->> Then you do the same vice versa. -->> There is a description in the manual. -->> -->> ----> -->> A B Is a circle. -->> <---- --> -->I have this set up, but I also have other slaves off of A like this: --> -->C --> -->^ -->| -->| --> -->A <----> B --> -->| -->| -->V --> -->D --> -->(note that C and D are simple slaves of A). --> -->The problem is that if you update B, it never reaches C and D unless you -->enable log-slave-updates, but then A and B seem to get into a loop so if -->you do an update on A, it updates on B which then send the update back to -->A, etc. I thought that MySQL wasn't supposed to do this (they have -->different server-id's), but it did. I would love to get this to work.
Sounds like a bug. If a query originated from A and ends back at A it should stop and not get inserted into A's binary log. There might be a server-id logic flaw that is allowing infinite replication loops. I would verify if this is the case (your bin-log should grow very fast with the same query but different server-ids) and write a bug report on it. -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
