any one of the servers could go down in many ways like 1) disk crash 2) replication failure 3) power failure 4) any hardware component failure 5) OS hang 6) Network failure 7) MYSQL bug 8) table corruption etc ....... 9) or just scheduled donwtime
in any case what really matters is the difference in status of the failed server from the point of failure to the point of recovery , so it all depends on how up-to-date your data is on the failed server.. here are some typical cases .. 1) you have everything intact ( data, bin logs, *.info files, etc ) then just bring back the failed server online and you are good to go 2) if you rebuild the server from scratch it depends on how current your backup is , in this take a snapshot of the main server ( either hot or cold depending on your setup) and start the multi-master replication .. In theory you really dont have to take the server offline unless it really demands , in the worst case as Barry said its lot easier to shut down the master (if you can afford some downtime) take a complete snapshot/backup/dump ( varies for MYISAM and INNODB) and bring back the failed server back online Hope this helps Kishore Jalleda On 3/28/06, 古雷 <[EMAIL PROTECTED]> wrote: > > Hello: > > If I use Multiple-Master Replication with two mysql server, when one of > them goes down(disk crashed) must I shutdown the good one to recover the > Multiple-Master Replication ? > > regards, > > gu lei