Eliminating database downtime in the presence of schema changes with ChronicDB
We are developing a solution for the problem of database downtime which is often incurred due to schema changes, and we are looking for feedback.
http://chronicdb.com/chronicdb_updates_100_live_database_connections_with_no_downtime
Releasing database changes with minimial downtime has been discussed before, from various viewpoints:
http://news.ycombinator.com/item?id=1313895 http://news.ycombinator.com/item?id=1265465 http://news.ycombinator.com/item?id=411282
And this suggests schema changes are also problematic in the Google App Engine. http://news.ycombinator.com/item?id=330413
Database schema changes seem to be a big problem. We would love to hear from anyone that has encountered this problem, and in situations where it has lead to downtime resulting in considerable costs. What was the cause of the downtime ? The duration ? The cost ? How often did such downtime occur ?
We also welcome feedback on the system design of the solution and suggestions for future improvements.
http://chronicdb.com/benefits_of_chronicdb
For anyone interested in joining, we are also looking for help.