HI, i have been developing a very complex app using couchdb for 2 years now. i'm trying at this point, to automate most of the things and foresee future problems in high demand, for the lack of human resources in the seed stage of this project.
i'm worried about 3 principal things: 1) when you delete a document, is supposed that revisions are only deleted after the cleanup, but the tombstones are still there, and it increase the amount of size exponentially. i found this post about it: http://eclipsesource.com/blogs/2015/04/20/how-to-finally-delete-documents-in-couchdb/ with a sort of solution, that i find tricky and insecure. i don't want be manipulating live DB's on live servers. any other option? 2) i'm not using replication and i will try to not use it (i'm all on amazon EC2 with EBS volumes, so the idea is to get bigger and not distributed for now). if you cleanup on source DB server, it will affect slaves too? 3)another thing that worries me, is the ID's generation. this statement is still true? http://docs.couchdb.org/en/latest/maintenance/performance.html#document-s-id There is so much difference between using the UUID generation and a 4 byte ID? any hints of how i can generate this 4 byte ID in nodejs in an easy way? i still need to check against the DB if the ID exists, with the 4 byte ID? Thanks for all the help ! :D
