database - Maintaining compatibility between different versions of software -
for our product (mostly written in java), have customers on multiple clusters. currently, in order deploy new version of software across our clusters take week do, isn't possible deploy clusters @ once , cannot achieved either (it's massive legacy codebase).
we're introducing new bit of functionality can push data across different clusters means potentially cause problems during our deployment window there customers on different versions of software @ point.
for example, if have 2 clusters 1 , 2 both own databases (sql server) table definitions identical. deploy latest version of software cluster 1 on day 1 , cluster 2 on day 2.
the latest version of software package includes new database field, data stored in field can pushed across clusters.
on day 1, customers on day 1 have latest version, attempt push data cluster 2 includes new field, cluster 2 database not have field, give exception.
what of approaches should considering maintain forward , backward compatibility during deployment window?
Comments
Post a Comment