We have a need to migrate a large number of 100GB databases onto a new infrastructure.
I would like to minimise downtime as much as possible, so am toying with the idea of migrating the clients to the new infrastructure on a clean,new, database and after the event load the existing data in behind the new data going in.
I fully understand that this is probably a "how long is a piece of string" question without seeing the schema (which is around 400 tables), but, does anyone know if this is a) possible and b) of any tools that might help.
These systems are all running on Ubuntu and are using mariadb.
I'm not sure what else I can provide to help extract a cogent answer, but I can but ask.
OK, so the stat of play is thus...
I have dumped the data from the tables I need to migrate and have added the --insert-ignore field to the dump to allow me to not get hung up on primary keys when I re-run things.
The data file produced is 23GB in total.
It all works well until I hit a table that has had the schema changed such that it has more columns than the source schema had at which point the inserts blow up because there is not enough data in the import line (the new fields do allow nulls, but I guess don't have a null default).
The error returned is Column count doesn't match value count on the table in question.
Any ideas on how to get around that?