John, Thank you for your reply. I have read the replication chapter and this seems to be easier, but I do have a few questions.
1. I am using InnoDB tables, at least there are innodb argument to mysqld in the start script, so should I use mysqldump instead of tar-ing the data dir to create the baseline for the slave? I believe thats what the docs are stating, but I wanted to be certain. 2. I want to have a failover scenario here, so what would the impact be on the mysql db if I go master->slave->master? 3. What happens if the master fails, dies, and goes offline. Must I change all my apps to connect to the IP of the slave for connections to work? I am unclear if replication will provide failover capabilities. 4. What is necessary to bring the master back up after a failure? Should I? or should I leave the slave (new master) up, and make the old master the new slave? 5. I am running 3.23.54 and I know I should upgrade, but its not going to happen today, so are there any show stopping bugs with this version? Thank you for your time, Andrew On Tue, 2003-12-02 at 17:10, John Griffin wrote: > Hi Andrew, > > I am not a Guru. I would suggest that you look at MySQL's excellent replication > facility rather than NFS mount a drive. Having your data on an NFS mounted drive > will significantly degrade the performance of your database. Replication will not. > > > John Griffin > > --Original Message-- > From: Andrew Hall [mailto:[EMAIL PROTECTED] > Sent: Tuesday, December 02, 2003 4:51 PM > To: [EMAIL PROTECTED] > Subject: mysql disaster recovery > > > Greetings Gurus, > > I have a mysql server that I need to create a disaster recovery system > for. What I am planning on doing is putting the data dir on a NFS > mounted directory so that I can start mysql on either of two servers in > case one dies. The inbound connections would be load balanced in a fail > over scenario, so the IP that clients will connect to will be on the > load balancer. > > I'm wondering if there is anything already developed that would test > mysql on the primary server, and if its not functioning, kill any > remaining mysql processes if necessary, and start it on the secondary. > This logic seems to be the biggest problem. > > Any suggestions, or other methodologies to implement this would be > welcome. > > Thank you for your time in advance, > > Andrew > -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
