Description
It is observed from several customers that replication failed to stop and restart after network issues. The root cause has been identified to be the calls to go-couchbase Connect() API, which do not have a timeout setting and could hang when network issues occur.
In order to fix this issue, the calls to the go-couchbase API have been replaced with direct http/https calls to target, which have timeout configured. This not only fixes the hanging problem, but also make the calls to target much cheaper.