I have a script that starts and stops replications on a very regular basis. Occasionally I'll make a request to cancel a replication as follows:
curl -H 'Content-Type: application/json' -X POST http://admin:***@***:5984/_replicate -d ' {"replication_id": "f509e26d1106d7022ae698c9a37e3c57+continuous+create_target", "cancel": true} ' and the request hangs - sometimes for a few minutes, occasionally for as long as 30 minutes. I can reproduce this on a regular basis on my production environment when juggling 30-40 replications between two CouchDB 1.6.1 servers in separate data centers. Any tips on how to troubleshoot this? Nick
