> (e) If you had just a one way master->slave relationship and you delete > data on the slave, that is very bad. The slave is now not the same as > the master, the record will not be re-inserted, if you go to update the > record on the master, when the command gets replicated to the slave an > error will generat and replication will halt awaiting you to manually > fix it. However, if you have setup a master/slave pair where > replication goes both ways as described in (a) or a ring as described in > (c) then you have no issue as all servers will be kept consistent.
Does anyone have scripts for checking the integrity of their slave servers? I've been writing some pretty in depth stuff in perl and I plan on releasing it eventually. It's pretty ugly right, even for perl. I will eventually clean it up. Some of the tests I do are: - make sure the same tables exist on the master and slaves - compare the schema (columns defs, keys, table types) - check that the row counts match - randomly check a percentage of rows in each table for large tables - for 'small' tables perform the checksumming as found in Sys Admin Mag's "Taming the Distributed Database Problem: A Case Study Using MySQL" (http://samag.com/articles/2004/0408/) Has anyone already gone through this effort already? -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
