On Tue, Jun 22, 2010 at 16:37, Robert Newson <[email protected]> wrote: > I'm sure there's a reason ("%% funky algorithm to preserve backwards > compatibility") but I wonder if this could be re-examined? It worries > me that testing multiple couchdb instances locally behaves differently > (actually, incorrectly) than testing multiple couchdb instances on > real, separate hosts.
Nevermind "real, separate hosts". There's no reason one shouldn't be able to run multiple couches on the same host and not confuse the replicator! It will complicate the logic a bit, but I think we can keep it fairly clean with an make_replication_id_old function or some such thing. If the new id (including port) isn't found, the old one can be checked to load the history, and saved under the new name going forward. This process ensures that old checkpoints aren't lost, though old couch instances that replicate with new ones will not see the new checkpoints. -Randall
