On Wed, Mar 2, 2011 at 1:29 PM, Filipe David Manana <[email protected]> wrote: > On Wed, Mar 2, 2011 at 9:09 PM, kowsik <[email protected]> wrote: >> My gut tells me the second case will be better for performance since >> the TCP connection (for the _changes feed) from A to B persists and >> the docs are sent one way without making a request (and hence less >> data transmitted per change). Is my hypothesis correct? I haven't yet >> looked at the packets to understand the differences between the two >> scenarios. But maybe someone else has already. > > This scenario changes with the new replicator (in trunk only) which > adds more parallelism and the difference between pull and push > replications is much shorter now. > > Recently Wayne Conrad tried out this new replicator and saw much > better performance/behaviour compared to the old one: > > http://mail-archives.apache.org/mod_mbox/couchdb-user/201103.mbox/%[email protected]%3E
Awesome! I have a follow up question on replication. I'm using couch as a message bus so have a lot of short-lived documents that come and go. Based on a previous thread, my understanding is that these documents are kept around in a deleted state forever for conflict resolution. If OTOH, I purge these documents, do the purge changes get propagated through replication? Also is there a way in _all_docs to walk just the deleted documents so that I can purge them? Thanks, K. --- http://twitter.com/pcapr http://www.pcapr.net http://labs.mudynamics.com
