So I spent the last few days occasionally tackling the problem of why I could not get my hot-failovers slave-io thread working on ubuntu server ... then I compared my masters config files and noticed that log_bin in the non working master and log-bin in the working master ... so then I reset master; redumped the database with --master-data and reloaded the slave .. then start slave and ...... whala! it worked ... another nice feature is that I only had to 'lock' our master DB for about 9 minutes while the dump was happening the prior method of copying the the db files between servers took about an hour of the db's being locked ... typically not noticed in the wee hours of the morning but 9 minutes of locked tables is much better .. and if things go right it will be two or three years before we need to do it again

Create Discussion