I am wondering: given that migrations often involve schema changes and subsequent data inserts of updates (e.g. update on a column just added by 'alter table'), what are your experiences with the fact that C* does not provide a way to control that the schema change actually occurred on all nodes before the update hits the cluster.
Do you have any pointers to discussions with the community that suggests that such migrations are even possible in a predictable fashion?
Hi,
I am wondering: given that migrations often involve schema changes and subsequent data inserts of updates (e.g. update on a column just added by 'alter table'), what are your experiences with the fact that C* does not provide a way to control that the schema change actually occurred on all nodes before the update hits the cluster.
Do you have any pointers to discussions with the community that suggests that such migrations are even possible in a predictable fashion?
Jan