That's pretty much it, indeed. You need to make absolutely sure that no more connections can be made to the old, broken master, though - even if you have to physically pull the network or power cable. Failover services refer to this as STONITH: Shoot The Other Node In The Head.
Don't think "but it went down anyway, that's why I failed over" - sometimes the service unexpectedly manages to recover itself after you've switched; and any data that gets inserted into the wrong system is as good as lost, unless you want to go through a manual reconciliation before reinitialising the replication. On Tue, Oct 19, 2010 at 11:55 AM, <a.sm...@ukgrid.net> wrote: > Quoting Machiel Richards <machi...@rdc.co.za>: > > >> The question I have however is how do you fail over to the slave >> server in the event that the master server is unavailable and then how >> to revert back to the master server once the server is available again. >> >> > Hi, > > to fail over to the slave, you dont need to do anything (except be sure > that replication is off). That is, tell your apps etc to use the slave > IP/hostname, change the slave server IP to that of the master or update DNS > so that all connections are made to the slave instead of the master. > > To reinstate the master after a failover, you must copy your live database > (now on slave) to the master and reinitiate replication pretty much as you > did when replication was initially setup. Copying the live DB can be > achieved via restoring a backup (from the live slave) or you could look at > Maakit which can analyse table differences and copy only changed data > (useful for very large DBs, Ive not personally used it tho). > > Thats about it I think, unless anyone wants to correct me on that... > > thanks Andy. > > > > > -- > MySQL General Mailing List > For list archives: http://lists.mysql.com/mysql > To unsubscribe: http://lists.mysql.com/mysql?unsub=vegiv...@tuxera.be > > -- Bier met grenadyn Is als mosterd by den wyn Sy die't drinkt, is eene kwezel Hy die't drinkt, is ras een ezel
