We have needed to do a couple of simple bug fixes to ignite proper, where there is no change to interfaces or internode communications. When we do this, we end up with these choices:
- Coordinate client and server code bases so that they are in lock step. Tedious with multiple clusters and test/dev versions. - Force the prior version number on the new builds, making it more tedious to understand what versions we are running. A standard practice would to ignore the last field in the version when doing a compatibility test, e.g., 2.5.0 and 2.5.foobar would be considered compatible. Is there some reason ignite requires and exact match? How do other Ignite users handle this problem? Thanks, -DH
